Slicing the AS/400 with Logical Partitioning: A

Transcription

Slicing the AS/400 with Logical Partitioning: A
Slicing the AS/400 with Logical Partitioning:
A How to Guide
Gottfried Schimunek, Danny Dupuche, Tim Fung, Paul Kirkdale,
Erik Myhra, Helmut Stein
International Technical Support Organization
http://www.redbooks.ibm.com
SG24-5439-00
SG24-5439-00
International Technical Support Organization
Slicing the AS/400 with Logical Partitioning:
A How to Guide
August 1999
Take Note!
Before using this information and the product it supports, be sure to read the general information in
Appendix F, “Special Notices” on page 205.
First Edition (August 1999)
This edition applies to Version 4, Release Number 4 of the AS/400 Operating System OS/400, Program
Number 5769-SS1
Note
This book is based on a pre-GA version of a product and may not apply when the product becomes
generally available. We recommend that you consult the product documentation or follow-on versions
of this redbook for more current information.
Comments may be addressed to:
IBM Corporation, International Technical Support Organization
Dept. JLU Building 107-2
3605 Highway 52N
Rochester, Minnesota 55901-7829
When you send information to IBM, you grant IBM a non-exclusive right to use or distribute the
information in any way it believes appropriate without incurring any obligation to you.
© Copyright International Business Machines Corporation 1999. All rights reserved
Note to U.S Government Users – Documentation related to restricted rights – Use, duplication or disclosure is
subject to restrictions set forth in GSA ADP Schedule Contract with IBM Corp.
Contents
Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .ix
Tables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
The Team That Wrote This Redbook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvi
Comments Welcome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xviii
Chapter 1. Logical Partitioning Concepts . . . . . . . . . .
1.1 Logical Partitioning History . . . . . . . . . . . . . . . . . . . .
1.2 Scenarios for Logical Partitioning . . . . . . . . . . . . . . .
1.2.1 Multiple Production Environments. . . . . . . . . . .
1.2.2 Mixed Production and Test Environments . . . . .
1.2.3 Implementing an Archiving Solution . . . . . . . . .
1.2.4 System Consolidation . . . . . . . . . . . . . . . . . . . .
1.2.5 Multidatabase Applications . . . . . . . . . . . . . . . .
1.2.6 Minimizing Backup and Recovery Windows . . .
1.2.7 Consolidating Worldwide Operations . . . . . . . .
1.2.8 A Perfect Combination with Domino Clustering .
1.2.9 Three-Tiered Application Architecture. . . . . . . .
1.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4 Partitioning Terminology . . . . . . . . . . . . . . . . . . . . . .
1.5 Primary Partition. . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.6 Secondary Partitions . . . . . . . . . . . . . . . . . . . . . . . .
1.7 Conceptual View of Logical Partitioning . . . . . . . . . .
1.8 Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.9 Marketplace Positioning . . . . . . . . . . . . . . . . . . . . . .
1.10 What Makes the Difference . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. .1
. .1
. .2
. .2
. .3
. .4
. .5
. .6
. .6
. .7
. .9
. 10
. 11
. 11
. 12
. 12
. 12
. 13
. 15
. 15
Chapter 2. Planning Considerations . . . . . . . . . . . . . . . . . .
2.1 Understanding Customer Requirements . . . . . . . . . . . . . .
2.1.1 Single Points of Failure. . . . . . . . . . . . . . . . . . . . . . .
2.1.2 Availability Requirements . . . . . . . . . . . . . . . . . . . . .
2.1.3 Resource Requirements . . . . . . . . . . . . . . . . . . . . . .
2.1.4 Application Stability . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.5 Security Requirements . . . . . . . . . . . . . . . . . . . . . . .
2.2 Understanding the New Complexity of Your Environment .
2.3 Logical Partitioning Configuration and Ordering Process .
2.4 Hardware Considerations . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.1 Supported AS/400e Models . . . . . . . . . . . . . . . . . . .
2.5 Software Considerations . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 17
. 17
. 17
. 18
. 19
. 19
. 20
. 20
. 21
. 23
. 23
. 24
© Copyright IBM Corp. 1999
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
iii
2.5.1 Licensed Program Products — Licensing . . . . . . . . .
2.5.2 Software Problems That Can Affect All Partitions . . .
2.5.3 Installing V4R4M0 . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.4 Application Dependencies between Partitions . . . . .
2.5.5 National Language Considerations . . . . . . . . . . . . . .
2.5.6 Future Release Co-Existence . . . . . . . . . . . . . . . . . .
2.5.7 Published Software Limits . . . . . . . . . . . . . . . . . . . .
2.6 Interface to Create and Manage Partitions . . . . . . . . . . . .
2.7 Deciding What Runs in Primary and Secondary Partitions
2.8 Choosing Bus- or IOP-Level Partitioning . . . . . . . . . . . . .
2.8.1 Bus-Level Partitions . . . . . . . . . . . . . . . . . . . . . . . . .
2.8.2 IOP-Level Partitions . . . . . . . . . . . . . . . . . . . . . . . . .
2.9 Moving to a Partitioned Environment . . . . . . . . . . . . . . . .
2.10 System Naming Convention . . . . . . . . . . . . . . . . . . . . . .
2.11 System Serial Number . . . . . . . . . . . . . . . . . . . . . . . . . .
2.12 Security . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.13 Updating Your Save Strategy . . . . . . . . . . . . . . . . . . . . .
2.14 Published Hardware Limits . . . . . . . . . . . . . . . . . . . . . . .
2.15 DASD Compression Support . . . . . . . . . . . . . . . . . . . . .
2.16 Dual Tape Attach Support to Multiple Partitions . . . . . . .
iv
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 24
. 25
. 25
. 26
. 26
. 26
. 26
. 26
. 28
. 29
. 30
. 30
. 31
. 31
. 32
. 32
. 33
. 37
. 37
. 37
Chapter 3. Operational Considerations . . . . . . . . . . . . . . . . . . . . .
3.1 Installing New Hardware Ready for Partitioning . . . . . . . . . . . . . .
3.2 Temporarily Making New Hardware Unavailable. . . . . . . . . . . . . .
3.2.1 Processors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.2 Main Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.3 Interactive Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.4 I/O Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3 Moving Existing IOPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4 Retrieving the Resources Allocated to a Partition . . . . . . . . . . . . .
3.5 Changing Partition Processing Resources . . . . . . . . . . . . . . . . . .
3.6 Saving and Restoring Partitions . . . . . . . . . . . . . . . . . . . . . . . . . .
3.6.1 Saving the Logical Partition Configuration . . . . . . . . . . . . . .
3.6.2 Saving Data from Logical Partitions . . . . . . . . . . . . . . . . . . .
3.6.3 Restoring Data to Logical Partitions . . . . . . . . . . . . . . . . . . .
3.7 System Values Affected by Logical Partitioning . . . . . . . . . . . . . .
3.7.1 QPRCMLTTSK — Processor Multi-Tasking . . . . . . . . . . . . .
3.7.2 QPWRRSTIPL — Automatic IPL after Power Restored. . . . .
3.7.3 QRMTIPL — Remote Power On and IPL . . . . . . . . . . . . . . .
3.7.4 QUPSDLYTIM — Uninterruptible Power Supply Delay Time .
3.7.5 Other Related System Values and Commands . . . . . . . . . . .
3.8 Security Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.9 Date and Time Processing between Partitions . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 39
. 39
. 40
. 40
. 41
. 42
. 43
. 43
. 44
. 44
. 45
. 46
. 46
. 46
. 47
. 47
. 47
. 48
. 48
. 49
. 51
. 51
Slicing the AS/400 with Logical Partitioning: A How to Guide
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
3.10 Main Storage Dump Processing . . . . . . . . . . . . . . . . . . . . . .
3.11 Managing the Hardware Resource Configuration . . . . . . . . .
3.11.1 Hardware Resource Information from DST and SST . . .
3.11.2 Hardware Resource Information from Logical Partitions
3.11.3 Removing Hardware Resources from Logical Partitions
3.11.4 Printing the System Configuration. . . . . . . . . . . . . . . . .
3.12 IPL Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.13 Licensed Key Management . . . . . . . . . . . . . . . . . . . . . . . . .
3.14 Work Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.14.1 Machine Pool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.14.2 QPFRADJ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 52
. 53
. 53
. 55
. 57
. 58
. 59
. 59
. 60
. 60
. 61
Chapter 4. Change Management for a Logical Partition .
4.1 Managing Partitions . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2 Managing User Applications in a Logical Partition . . . . .
4.2.1 Distributing Changes through Management Central
4.3 Program Temporary Fix (PTF) Management . . . . . . . . .
4.3.1 Using Management Central in PTF Distribution . . .
4.4 Managing Security in Logical Partition . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 63
. 63
. 63
. 64
. 65
. 66
. 68
Chapter 5. Problem Management for a Logical Partition . . . .
5.1 Problem Determination for a Logical Partition . . . . . . . . . . . .
5.2 Reporting to IBM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3 System Reference Codes (SRC) . . . . . . . . . . . . . . . . . . . . .
5.3.1 Licensed Internal Code Log . . . . . . . . . . . . . . . . . . . . .
5.4 Power Interruptions in the Logical Partition . . . . . . . . . . . . .
5.5 Electronic Customer Support Lines . . . . . . . . . . . . . . . . . . .
5.6 Logical Partition Configurations . . . . . . . . . . . . . . . . . . . . . .
5.6.1 Working with Configuration Data for Logical Partitions .
5.7 Recovery for Logical Partitions . . . . . . . . . . . . . . . . . . . . . . .
5.7.1 Recovery Considerations with Logical Partitions. . . . . .
5.7.2 Information on Recovering after System Initialization . .
5.7.3 Restoring User Data to a Logical Partition . . . . . . . . . .
5.7.4 Recovering User Data from a Deleted Logical Partition
5.8 Service Provider Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 69
. 69
. 70
. 70
. 72
. 72
. 72
. 73
. 73
. 77
. 77
. 78
. 78
. 78
. 79
Chapter 6. Capacity Planning and Performance . . . . .
6.1 Determining Customer Requirements . . . . . . . . . . . .
6.2 Commercial Processing Workload (CPW) Concepts .
6.2.1 AS/400 Workload — Technical View. . . . . . . . .
6.2.2 Different CPW Values . . . . . . . . . . . . . . . . . . . .
6.2.3 AS/400 System Models . . . . . . . . . . . . . . . . . .
6.2.4 New AS/400e server 7XX . . . . . . . . . . . . . . . . .
6.2.5 AS/400 Server Models . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 81
. 81
. 82
. 82
. 83
. 83
. 84
. 86
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
v
vi
6.2.6 AS/400 Custom Server Models . . . . . . . . . . . . . . . . . . . . . . .
6.3 Determining Batch CPW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3.1 BATCH400 — Enabling the AS/400 Batch Window. . . . . . . .
6.3.2 Batch Capacity Planning for LPAR . . . . . . . . . . . . . . . . . . . .
6.4 Determining Partition Processing Resources . . . . . . . . . . . . . . . .
6.4.1 Determining the Number of Processors . . . . . . . . . . . . . . . .
6.4.2 Determining Main Storage . . . . . . . . . . . . . . . . . . . . . . . . . .
6.4.3 Determining Interactive Performance . . . . . . . . . . . . . . . . . .
6.5 OptiConnect/400 Performance Considerations . . . . . . . . . . . . . . .
6.6 Available Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.6.1 Performance Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.6.2 BEST/1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.7 Creating a New Hardware Feature in BEST/1. . . . . . . . . . . . . . . .
6.7.1 Going to the Partition Processing Configuration . . . . . . . . . .
6.7.2 Calculating New Performance Ratings for Logical Partitions .
6.7.3 Creating the New Hardware Feature in BEST/1 . . . . . . . . . .
6.7.4 Using the New Hardware Feature . . . . . . . . . . . . . . . . . . . . .
6.7.5 Performance Management/400. . . . . . . . . . . . . . . . . . . . . . .
6.8 Performance Management Using Management Central . . . . . . . .
6.8.1 Setting Up Your Central System . . . . . . . . . . . . . . . . . . . . . .
6.8.2 Collecting Performance Data . . . . . . . . . . . . . . . . . . . . . . . .
6.9 Monitoring Realtime System Performance . . . . . . . . . . . . . . . . . .
6.9.1 Management Central . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.9.2 Running a Monitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.10 On-Going Performance Monitoring . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. . 87
. . 88
. . 89
. . 89
. . 92
. . 93
. . 95
. . 96
. . 97
. . 98
. . 99
. . 99
. . 99
. 100
. 102
. 103
. 108
. 109
. 110
. 110
. 113
. 118
. 118
. 128
. 130
Chapter 7. Inter-Partition Communication . . . . . . . . . .
7.1 OptiConnect/400 Capabilities . . . . . . . . . . . . . . . . . . .
7.1.1 OptiConnect/400 . . . . . . . . . . . . . . . . . . . . . . . .
7.1.2 OptiMover for OS/400 . . . . . . . . . . . . . . . . . . . . .
7.2 OS/400 Functions Using OptiConnect/400 . . . . . . . . .
7.2.1 Distributed Data Management (DDM) . . . . . . . . .
7.2.2 ObjectConnect/400 . . . . . . . . . . . . . . . . . . . . . . .
7.2.3 Remote Journal Function . . . . . . . . . . . . . . . . . .
7.3 Virtual OptiConnect/400 . . . . . . . . . . . . . . . . . . . . . . .
7.3.1 Transferring Data between Logical Partitions . . .
7.3.2 Setting Up Virtual OptiConnect over SNA . . . . . .
7.3.3 Starting Virtual OptiConnect . . . . . . . . . . . . . . . .
7.3.4 Ending Virtual OptiConnect . . . . . . . . . . . . . . . .
7.3.5 Determining Virtual OptiConnect Activity . . . . . .
7.3.6 Setting Up OptiConnect/400 over TCP/IP . . . . . .
7.4 Application Examples Using Virtual OptiConnect/400 .
7.4.1 SNA Distribution Services (SNADS) . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 133
. 133
. 133
. 135
. 135
. 135
. 137
. 138
. 139
. 140
. 141
. 148
. 148
. 149
. 152
. 158
. 158
Slicing the AS/400 with Logical Partitioning: A How to Guide
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
7.4.2
7.4.3
7.4.4
7.4.5
7.4.6
7.4.7
7.4.8
Remote Journal Function . . . . . . . . . . . . . . . . . . . . . . .
Data Replication Using Journals . . . . . . . . . . . . . . . . . .
Data Replication Using the Remote Journal Function . .
Setting Up the Remote Journal Function . . . . . . . . . . .
Benefits of the Remote Journal Function . . . . . . . . . . .
Using Your Database 24 Hours a Day, 7 Days a Week.
DB2 Multisystem. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
..
..
..
..
..
..
..
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. 159
. 160
. 161
. 162
. 164
. 164
. 165
Appendix A. Logical Partitioning Terminology . . . . . . . . . . . . . . . . . . . 167
A.1 Processor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
A.2 Main Storage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
A.3 Bus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
A.4 Input/Output Processor (IOP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
A.5 Multifunction IOP (MFIOP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
A.6 Load Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
A.7 System Licensed Internal Code (SLIC). . . . . . . . . . . . . . . . . . . . . . . . . . 168
A.8 Partition Licensed Internal Code (PLIC) . . . . . . . . . . . . . . . . . . . . . . . . . 168
A.9 Hypervisor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
A.10 Removable Media Device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
A.11 Console . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
A.12 Expansion Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
A.13 Ownership and Sharing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
A.14 Addition and Removal. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
A.15 OptiConnect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
Appendix B. Sample RPG and C Programs . . . . . . . . . . . . . . . . . . . . . . 171
B.1 Sample ILE RPG Program — Output to Spooled File . . . . . . . . . . . . . . 171
B.2 Sample C Program — Output to Screen. . . . . . . . . . . . . . . . . . . . . . . . . 173
Appendix C. Accessing LPAR Functions . . . . . . . . . . . . . . . . . . . . . . . . 175
Appendix D. Operations Console Connection for Partitioned AS/400177
D.1 What Operations Console Is . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
D.2 Direct-Attached Operations Console in Secondary Partition . . . . . . . . . 178
Appendix E. Interactive and Batch CPW Determination . . . . . . . . . . . 189
E.1 Overview of the Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
E.2 Interactive Performance Determination . . . . . . . . . . . . . . . . . . . . . . . . . 190
E.3 Batch Performance Determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
Appendix F. Special Notices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Appendix G. Related Publications. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
G.1 International Technical Support Organization Publications . . . . . . . . . . 209
vii
G.2 Redbooks on CD-ROMs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
G.3 Other Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
How to Get ITSO Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
IBM Redbook Fax Order Form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
ITSO Redbook Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
viii
Slicing the AS/400 with Logical Partitioning: A How to Guide
Figures
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
Chronology of Logical Partitioning Implementation . . . . . . . . . . . . . . . . . . . 1
Mixed Production and Test Environments . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Setting Up an Archiving Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Minimizing Backup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Different Time Zones on One AS/400 System. . . . . . . . . . . . . . . . . . . . . . . 8
Logical Partitions for Different Time Zones . . . . . . . . . . . . . . . . . . . . . . . . . 9
Lotus Domino Clustering over AS/400 Logical Partitions . . . . . . . . . . . . . 10
AS/400 Logical Partitioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Generic View of Logical Partitioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Clustered Configuration with a Partitioned Node . . . . . . . . . . . . . . . . . . . . 14
Replicating Data from One Partition to Another. . . . . . . . . . . . . . . . . . . . . 19
Using Multiple Tape Drives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Saving from a Second Copy of the Data . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Using ObjectConnect to Move Data to Another Partition . . . . . . . . . . . . . . 36
Change Partition Processing Resources — Processors . . . . . . . . . . . . . . 41
Change Partition Processing Resources — Main Storage . . . . . . . . . . . . 42
Change Partition Processing Resources — Interactive Performance . . . . 43
Work with Partition Configuration Display with Changes Pending. . . . . . . 45
Work with Partition Status Display . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Display Secondary Partition Reference Code History . . . . . . . . . . . . . . . . 52
Hardware Service Manager — Initial Menu . . . . . . . . . . . . . . . . . . . . . . . . 54
Hardware Service Manager — Sample Resources Display . . . . . . . . . . . 55
Display Allocated I/O Resources from Primary Partition . . . . . . . . . . . . . . 56
Display System I/O Resources Display . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Remove I/O Resources Display. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Sending a Package in Management Central . . . . . . . . . . . . . . . . . . . . . . . 65
Sending Fixes Using Management Central . . . . . . . . . . . . . . . . . . . . . . . . 67
Display Secondary Partition Reference Code History . . . . . . . . . . . . . . . . 71
Include Reference Code Detail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Recover Configuration Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
System Properly Labeled. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
Interactive CPW Utilization — Model 600 . . . . . . . . . . . . . . . . . . . . . . . . . 84
Interactive CPW Utilization — Model 730 . . . . . . . . . . . . . . . . . . . . . . . . . 85
Interactive CPW Utilization — Model S30 Processor #2259 . . . . . . . . . . . 86
Interactive CPW Utilization — Model S20 . . . . . . . . . . . . . . . . . . . . . . . . . 87
Assigning CPW for Batch. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Creating Logical Partitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
Going to the Partition Processing Configuration (Part 1 of 4) . . . . . . . . . 100
Going to the Partition Processing Configuration (Part 2 of 4) . . . . . . . . . 101
Going to the Partition Processing Configuration (Part 3 of 4) . . . . . . . . . 101
ix
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
x
Going to the Partition Processing Configuration (Part 4 of 4) . . . . . . . . . 102
Work with CPU Models — BEST/1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Copy CPU Model — BEST/1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Change CPU Model — BEST/1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Save Hardware Characteristics — BEST/1 . . . . . . . . . . . . . . . . . . . . . . . 106
Work with Members Using PDM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Copy Members. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Change CPU and Other Resource Values . . . . . . . . . . . . . . . . . . . . . . . 108
Hardware Characteristics Menu. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Adding an Endpoint System in Management Central . . . . . . . . . . . . . . . 111
Creating New System Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
Start Collecting Performance Data on an Endpoint System . . . . . . . . . . 114
Start Collection Services on a System Group . . . . . . . . . . . . . . . . . . . . . 116
Creating Database Files from Collection Services. . . . . . . . . . . . . . . . . . 117
Adding a New Monitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Edit Metric . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Treshold Trigger Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Edit Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
New Command Definition for Thresholds . . . . . . . . . . . . . . . . . . . . . . . . 127
Performance Monitor — CPU Utilization and Disk Storage . . . . . . . . . . . 128
Monitoring Two Logical Partitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
OptiConnect/400 Concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
Distributed Data Management Concept. . . . . . . . . . . . . . . . . . . . . . . . . . 136
Virtual OptiConnect/400 Concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
Read Request Using Virtual OptiConnect . . . . . . . . . . . . . . . . . . . . . . . . 141
Creating a DDM File to Use OptiConnect (Part 1 of 2) . . . . . . . . . . . . . . 143
Creating a DDM File to Use OptiConnect (Part 2 of 2) . . . . . . . . . . . . . . 143
Matching Parameters for Extended Function Path Routing. . . . . . . . . . . 144
Create Controller Description on LPAR1 to Connect to LPAR2 . . . . . . . 145
Create Controller Description on LPAR2 to Connect to LPAR1 . . . . . . . 145
Create Device Description on LPAR1 to Attach LPAR2CTL . . . . . . . . . . 146
Create Device Description on LPAR2 to Attach LPAR1CTL . . . . . . . . . . 147
Work with Active Jobs — QSOC Subsystem. . . . . . . . . . . . . . . . . . . . . . 150
Work with OptiConnect Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
Configuration of the OptiConnect Interface — LAN Version . . . . . . . . . . 153
ADDTCPIFC — Add TCP/IP Interface for OptiConnect. . . . . . . . . . . . . . 155
Configuring OptiConnect Interface (Part 1 of 2). . . . . . . . . . . . . . . . . . . . 156
Configuring OptiConnect Interface (Part 2 of 2). . . . . . . . . . . . . . . . . . . . 156
Hot-Backup Environment without Remote Journal Function . . . . . . . . . . 161
Hot-Backup with Remote Journal Function . . . . . . . . . . . . . . . . . . . . . . . 162
Add RDB Directory Entry — SNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
Change DDM TCP/IP Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
Using Remote Journal Function for 24-by-7 Database Availability . . . . . 165
Slicing the AS/400 with Logical Partitioning: A How to Guide
84. Distribution of Database Files across Logical Partitions . . . . . . . . . . . . . 166
85. Operations Console Attachment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
86. AS/400 Operations Console Modem . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
87. Select Console IOP with Filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
88. Select Console IOP without the Filtering . . . . . . . . . . . . . . . . . . . . . . . . . 182
89. Work with Partition Configuration Menu. . . . . . . . . . . . . . . . . . . . . . . . . . 183
90. Select Default Electronic Customer Support Resource Menu . . . . . . . . . 183
91. Operations Console DST Sign On Window . . . . . . . . . . . . . . . . . . . . . . . 184
92. IPL or Install the System Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
93. DST Main Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
94. Work with DST Environment Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
95. Work with System Devices Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
96. Select Console Type Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
97. Choosing Option 5 to Work with the Model INTER01 . . . . . . . . . . . . . . . 191
98. Choosing Option 1 to Work with the Workloads of the Model INTER01 . 191
99. Choosing Option 8 to Select the Required Workloads. . . . . . . . . . . . . . . 192
100.Save Workload to Workload Member . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
101.Work with Workloads. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
102.Work with Best/1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
103.Work with Workloads. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
104.Add Saved Workload. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
105.Work with Workload . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
106.User-Defined AS/400 Partition Model — 650-LP04 . . . . . . . . . . . . . . . . 197
107.CPU Utilization versus Time Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
108.BEST/1 Projection for the Batch Processing . . . . . . . . . . . . . . . . . . . . . . 199
109.Workload Profile of M53S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
110.Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
111.Select CPU . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
112.Projected Batch Performance for the Combined Workload. . . . . . . . . . . 202
xi
xii
Slicing the AS/400 with Logical Partitioning: A How to Guide
Tables
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Logical Partitioning Marketplace Offerings as of February 1999 . . . . . . . . 15
Supported Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
AS/400 Model 600 — Feature Processor #2136 . . . . . . . . . . . . . . . . . . . . 84
AS/400 Model 730 — Feature Processor #2068 . . . . . . . . . . . . . . . . . . . 85
AS/400 Model S30 — Feature Processor #2259. . . . . . . . . . . . . . . . . . . . 86
AS/400 Model S20 — Feature Processor #2170. . . . . . . . . . . . . . . . . . . . 87
Systems Supporting LPAR — Batch CPW . . . . . . . . . . . . . . . . . . . . . . . . 90
Assigning CPW for a Batch Environment . . . . . . . . . . . . . . . . . . . . . . . . . 91
CPW Ratings for 7XX Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Capacity Planning Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
First Cut into Logical Partitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
Working Out the Number of Processors . . . . . . . . . . . . . . . . . . . . . . . . . . 94
Working Out Memory Allocation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Working Out Interactive Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
Metrics for Realtime Performance Monitor Graphs . . . . . . . . . . . . . . . . . 118
Threshold Trigger Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Logical Partitions Menu Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
xiii
xiv
Slicing the AS/400 with Logical Partitioning: A How to Guide
Preface
Take a tour of logical partitioning on the AS/400! Logical partitioning allows
one physical system to be divided into multiple logical partitions, each
endowed with its own resources and capable of running independently. This
redbook discusses the logical partitioning revolution on the AS/400 system. It
is designed for all levels of understanding. It targets the person who simply
needs an overall view of logical partitioning on the AS/400 system. And, it
targets the technical people who will be actively engaged in devising
solutions that involve logical partitioning, creating partitioned machines, and
managing and operating them.
A shift to a more consolidated view of computer systems has triggered a
renewed increase in machine sizes capable of consolidating multiple
environments. The AS/400 system has embraced this trend by increasing its
size and implementing logical partitioning, the concept of independent logical
machines within large physical machines. This redbook is about the logical
partitioning revolution on the AS/400 system. It is a new weapon in the
AS/400 system’s already rich armory.
Logical partitioning dates back to 1976, when IBM experimented with the
S/370. Other hardware manufacturers followed this IBM path with various
flavors of the concept. By 1997, companies such as Amdahl, Hitachi, and
SUN, along with IBM, had their own versions of system partitioning. In 1999,
the AS/400 system burst onto the scene with its own implementation.
Here is a brief guide to the chapters of this redbook:
• Chapter 1 explains the basic concepts of logical partitioning and provides
several scenarios where this new functionality can be beneficial.
• Chapter 2 offers practical information for the reader who needs to plan and
implement a logical planning solution. This chapter has numerous
references to other existing publications for finer detail as required.
• Chapter 3 considers the aftermath of logical partitioning and addresses
itself mainly to the day-to-day routine of working with logical partitioning.
• Chapters 4 and 5 extend Chapter 3 and deal with change management
and problem management in the context of logical partitioning.
• Chapter 6 discusses capacity planning and performance. It provides
valuable insight in the task of initially assessing the suitability of
implementing logical partitioning in a given situation, developing the
logical partitioning configuration to fit the situation, and ensuring the
continued optimal performance of the installation.
© Copyright IBM Corp. 1999
xv
• Chapter 7 explains the crucial topic of high-speed communication between
logical partitions. This is one of the many remarkable features of logical
partitioning.
• The redbook contains appendices and references to other publications.
This material is also beneficial to IBM customers, IBM Business Partners, and
IBM technical staff. Regardless of your level of expertise, this redbook tells
you all you need to know about logical partitioning on the AS/400 system.
The Team That Wrote This Redbook
This redbook was produced by a team of specialists from around the world
working at the International Technical Support Organization Rochester
Center.
Gottfried Schimunek is a certified IT Architect at the
IBM International Technical Support Organization,
Rochester Center. He writes extensively and teaches
IBM classes worldwide on all areas of application
development and performance. Before joining the ITSO
in 1997, Gottfried worked in the Technical Support
Center in Germany as a consultant for IBM Business
Partners and customers as well as for the IBM sales
force.
Danny Dupuche is a Senior Technical Consultant with
KAZ Computers Services, a major IBM Business Partner
in Australia. Originally from a mainframe background,
Danny started working with the S/38 in the early years of
that system. His areas of expertise are in capacity
planning and sizing, systems management, and high
availability solutions for the company’s customers. He
currently monitors and analyzes customer business
requirements and, in conjunction with the company’s
marketing arm, provides solutions to meet those goals.
xvi
Slicing the AS/400 with Logical Partitioning: A How to Guide
Tim Fung is a Consulting I/T Specialist at IBM
China/Hong Kong Limited. He has over ten years of
experience in the S/36, S/38 and AS/400. He works with
the AS/400 in IBM Hong Kong providing technical
marketing support to AS/400 sales representatives, IBM
Business Partners and customers. His areas of
expertise include availability solutions, capacity
planning, and the Internet.
Paul Kirkdale is a certified AS/400 IT Consultant
working for IBM in the UK. He has 10 years of experince
in working with the AS/400. Currently, he provides
technical support for high availability and clustering to
large AS/400 users in the UK and across Europe. Prior
to this role, he worked in the areas of application
development, database design, performance, and
systems management. Before joining IBM in 1990, Paul
worked for an IBM Business Partner where he provided
AS/400 and System/38 support to customers.
Erik Myrha is a Senior Advisory SE at the Customer
Support Center in Norway. His areas of expertise include
application development and performance. He has
worked at IBM for 22 years, including three years in the
Rochester ITSO.
Helmut Stein is an AS/400 IT Consultant working for
IBM in Germany. He has 23 years of experience in the
S/34, S/36, S/38, and AS/400 systems. He provides
technical pre-sales support to large AS/400 customers,
IBM sales representatives, and IBM Business Partners.
His areas of expertise include performance,
high-availability, server consolidation, and OptiConnect.
A special thanks goes to Joel Nelson from the Rochester Product Field
Support Center. He was very instrumental in finding answers related to
xvii
AS/400 hardware configurations and the use of Hardware Service Manager,
as well as the dedicated service tools.
Thanks goes also to the following people for their invaluable contributions to
this project:
Doug Anderson
Bill Armstrong
Sue Baker
Dan Birkestrand
Lynn Chung
Amit Dave
Selwyn Dickey
Mike Faunce
Ian Jarman
Mike Leslie
Lynn McMahon
Mark Manges
Naresh Nayar
John Reichel
Debbie Saugen
Jeff Scheel
Brenda Thompson
Garrett Winn
Comments Welcome
Your comments are important to us!
We want our redbooks to be as helpful as possible. Please send us your
comments about this or other redbooks in one of the following ways:
• Fax the evaluation form found in “ITSO Redbook Evaluation” on page 219
to the fax number shown on the form.
• Use the online evaluation form found at: http://www.redbooks.ibm.com/
• Send your comments in an Internet note to: [email protected]
xviii
Slicing the AS/400 with Logical Partitioning: A How to Guide
Chapter 1. Logical Partitioning Concepts
This chapter explains the basic concepts of AS/400 Logical Partitioning. The
end objective of logical partitioning (LPAR), released with OS/400 V4R4, is to
provide users with the ability to split a single AS/400 system into several
independent systems capable of running applications in multiple,
independent environments simultaneously. For example, logical partitioning
makes it possible for a user to run a single application using different sets of
data on separate partitions, as if it was running independently on separate
physical AS/400 systems.
1.1 Logical Partitioning History
IBM began the study of logical partitioning for the S/370 architecture in
Poughkeepsie, New York, in 1976. The project proved that logical partitioning
was a workable concept, but the idea had not been extended to actual
product implementation. In November 1984, Amdahl Corporation announced
the 580/Multiple Domain Feature on the IBM S/370 mainframe. It was the first
logical partitioning product and it was favorably accepted by the market.
Then, the evolution of partitioning began (see Figure 1).
Date
1985
Vendor
Amdahl
Product Name
Multiple Domain Feature on
IBM Mainframe
1988
IBM
PR/SM Logical Partitions
on IBM Mainframe
1989
Hitachi
Multiple Logical Processor
Facility on IBM Mainframe
1997
SUN
Dynamic System Domains
on SUN UE 10000 server
1999
IBM
Logical Partitioning on
AS/400
Figure 1. Chronology of Logical Partitioning Implementation
As we can see, partitioning has been available on IBM mainframes for more
than a decade. Over that period, it has evolved from a predominantly physical
partitioning scheme, based on hardware boundaries, to one that allows for
© Copyright IBM Corp. 1999
1
virtual and shared resources, with dynamic load balancing. In today’s
marketplace, partitioning has become a requirement. All the major mainframe
players, including IBM, offer partitioning.
History repeats itself. The factors that have driven the evolution of mainframe
partitioning over the past decade are now at work in the server system arena.
Partitioning is fast becoming a necessity there too. It is estimated that by the
year 2001, all major players in the server marketplace may offer some degree
of partitioning.
The AS/400 system is delivering its own complete version of partitioning, right
now. Logical partitioning implementation on an AS/400 system is an
adaptation of the System 390 logical partitions concept, with flexible and
granular allocation of system resources. It also offers flexibility in allocating
interactive performance and high-speed internal communications between
partitions.
The current hardware offering delivers a maximum of 12 processors, which
allows a maximum of 12 logical partitions. The same standard tools used
across physical machines, for example SQL, DDM, DRDA, Remote
journaling, and DB2 Multisystem, are available for use between logical
partitions.
1.2 Scenarios for Logical Partitioning
The range of scenarios possible with logical partitioning is limited only by
your imagination. Let us select a few of the most promising ones and explore
the solutions offered by the application of logical partitioning to see what can
be achieved.
1.2.1 Multiple Production Environments
Let us first consider the case of a company that offers outsourcing services to
its customers. The economics of the case convinced the CEO that the
consolidation of all the company’s physical machines into a single
multipartitioned machine would deliver improved services to customers and
streamline day-to-day operations. The profit expectations look attractive.
This service provider can purchase a multiprocessor model 7XX with enough
resources to create several completely independent logical partitions to meet
customer requirements. These partitions can operate as independent
machines. For example, some partitions may run client/server workloads,
while others run purely interactive workloads, still others may run mixed
workloads. The permutations are endless.
2
Slicing the AS/400 with Logical Partitioning: A How to Guide
1.2.2 Mixed Production and Test Environments
Generally, production and test environments must be kept separate. Logical
partitioning achieves this objective in a simple way.
One or several partitions can be allocated for production only. Others are set
up to handle application testing, either to validate application changes or to
test applications under new releases of system software.
One special case today is testing applications for Year 2000 ready (Y2K)
compliance. A logical partition can be set up with its own special environment
having its own date, time, and system values. Communications between the
partitions can be set up as required. Figure 2 illustrates this concept.
Figure 2. Mixed Production and Test Environments
Without logical partitioning, the only practical way of testing Y2K would be to
purchase additional hardware and software. However, the disadvantage may
be that the hardware and software may not be required subsequently. Note
that system software is currently charged at the physical machine level.
Therefore, it can be run on several independent partitions, without an
additional charge.
Logical Partitioning Concepts
3
Partitioning provides a solution by making it possible to use an independent
partition for testing. When testing is over, the resources allocated to this
partition can be returned to the production partition or elsewhere as required.
1.2.3 Implementing an Archiving Solution
In this scenario, the customer has a very large database containing a mix of
live and historical data. Most of the interactive and batch jobs use the live
records on the database. However, the customer needs to access the
historical data both interactively and in batch mode from time to time, with
reasonable response times. Furthermore, they also need to reduce their
backup time, without jeopardizing data security. The parameters of this
situation are:
•
•
•
•
Database size........................500GB
Number of logical views............ 60
Live data.............................. 200GB
Historical data....................... 300GB
Assuming that the application can distinguish between live and historical
data, one possible solution is to create a three-partitioned configuration as
shown in Figure 3.
Figure 3. Setting Up an Archiving Solution
The outcome of this implementation would be:
4
Slicing the AS/400 with Logical Partitioning: A How to Guide
• Faster updates of live data because of reduced database index tables.
• Seamless access to historical data through high-speed inter-partition
communication.
• Reducing the size of the database also reduces the size of the database
index tables and, therefore, the amount of work the CPU has to do to
update these tables.
• Overall improved response times.
• Reduced backup requirements. Only live data needs to be backed up at
frequent intervals. Backing up archived data needs to be performed only
at the time when live data transitions to archived status.
• Report consolidation through DB2 Multisystem or Remote SQL over
high-speed inter-partition communication.
1.2.4 System Consolidation
Today's AS/400e server offers over 300 times the performance range from
the smallest to the largest servers. Many companies are taking advantage of
the high degree of scalability by consolidating existing AS/400e applications
onto fewer systems, often in regional data centers.
Still many companies are attracted to the concept of distributed computing, in
particular on the AS/400, Unix, and NT servers. The attraction of distributed
servers is due to the low initial investment costs and the flexibility of
deploying applications customized to the needs of specific departments.
However, the realities of managing distributed UNIX and NT platforms proved
far more difficult and costly than initially imagined.
In response, many companies are now investigating opportunities to
consolidate their decentralized and distributed systems to regain both
financial and administrative control of their IT resources. This trend in the
industry is referred to as server consolidation. In a recent International Data
Corporation (IDC) white paper, Server Consolidation: An AS/400 TCO
Analysis, 49% of respondents to their survey were in the process of
consolidating servers and storage.
With its strengths as both an enterprise and distributed server, the AS/400
system has a unique position as a server consolidation platform for AS/400
consolidation, Domino for AS/400, and Windows NT Integrated Netfinity
Server. With the support of logical partitions on the AS/400 system, server
consolidation becomes a key element of the AS/400 enterprise systems
future.
Logical Partitioning Concepts
5
IBM has a worldwide server consolidation solutions team to assist customers
that are considering server consolidation. IBM has a range of consolidation
solutions. They include a consistent methodology and a wide range of
services, from expert advice in planning to optimizing and benchmarking
consolidated applications. In addition to offering planning services, IBM
Global Services has experienced specialists who can perform an AS/400
consolidation.
For more information, see IBM’s Web site at: http://www.as400.ibm.com/sc
This Web site includes a wide range of information on server consolidation,
including reference to the following consultant reports:
• Getting It Together: AS/400e — New Ammunition for Server Consolidation
(D.H Andrews Group)
• Server Consolidation: An AS/400 TCO Analysis (International Data
Corporation)
1.2.5 Multidatabase Applications
Some businesses may require running the same application against
independent databases. It may be that, because of naming conventions or
other internal restrictions, the application cannot be run against multiple
independent databases on the same physical machine. The solution before
logical partitioning would be either to purchase additional hardware or modify
the application.
With logical partitioning, each instance of the application can be run within its
own independent partition. In addition, communications between the
independent partitions can be used for consolidated reporting if required.
1.2.6 Minimizing Backup and Recovery Windows
The imperatives driving today’s businesses demand total machine availability,
7 days-a-week, 24 hours-a-day (7 by 24). On the other hand, data security
demands that a foolproof backup strategy be in place to meet unforeseen
contingencies. A third ingredient in this situation is the relentless growth of
databases as businesses become more complex. Logical partitioning can
provide one solution to balance these conflicting needs.
Let us look at a scenario where backup is becoming so time consuming that
the production window is dwindling fast. Figure 4 illustrates a possible
solution to minimize backup of the production database. Incidentally, this
scenario also facilitates recovery.
6
Slicing the AS/400 with Logical Partitioning: A How to Guide
Figure 4. Minimizing Backup
In this scenario, production is running on partition 0 and all updates are
replicated on partition 1, using remote journaling. At preset intervals, the
partition 1 update is suspended, and the partition database is backed up.
After the backup, the partition 1 database is re-synchronized with partition 0,
by applying all accumulated journal entries from partition 0 to partition 1.
This scenario provides for recovery of the production database onto the same
partition or a different physical machine, with minimum inconvenience.
1.2.7 Consolidating Worldwide Operations
Time zones are a way of organizing time so that people living in a large area
(country or part of a continent) share the same time, just to make things easy.
People living at about the same longitude usually have about the same local
time. There are 24 standard time zones around the world. These are all an
integral offset of hours to Universal Time Coordinated (UTC)-time, for
example, +5 hours or -10 hours to UTC. If UTC-time is 12:00 (AM), then
people having offset +5 hours, have local time 17:00 (or 5:00 PM).
If your business needs different time zones on a single AS/400 system, there
are no AS/400 system values available to support you. If you, for example,
have a system located in Hong Kong and have offices located in the US, the
users signing on to that system from the US offices will always have the Hong
Kong time. Your application has to manage the different time zones for your
users, which may be located around the whole world (Figure 5). Even if you
can manage the different time zones by your application, you still need to find
Logical Partitioning Concepts
7
a way to manage your downtimes. Your system cannot be saved while your
users in the US are still running their interactive applications. You need to
have batch windows for running batch programs.
Figure 5. Different Time Zones on One AS/400 System
One way to provide different time zones to your users is to split your AS/400
system from a single system image to a system with logical partitions. Even if
the time is only provided by the service processor of the primary partition, you
can have different times on logical partitions. The logical partitions manage
the time difference to the primary partition.
Usually your application has only one database for all of your central and
distributed offices. You have to decide whether you want a database partition
and application partitions for all your different time zones or you want your
central site application and database in one partition.
8
Slicing the AS/400 with Logical Partitioning: A How to Guide
Figure 6. Logical Partitions for Different Time Zones
In Figure 6, the single AS/400 system image was split up into four partitions.
In the primary partition, you have the database for all secondary partitions.
For each time zone, you have your own logical partition. The application in
the logical partition has to be changed. Your database resides in the primary
partition, and the application in the logical partition needs to know where the
database is.
Using this method for different time zones on a partitioned system, you still
need to manage your downtime. There is still only one image of your
database, and you have to stop your applications while you backup your
database. If you need to provide 24-hour service for your users, you need to
mirror your database on a logical partition. Different software vendors provide
solutions for doing that. The backup is done only from the secondary
partition, and your users can work with the applications for 24 hours.
1.2.8 A Perfect Combination with Domino Clustering
E-mail and groupware software, such as Lotus Domino for AS/400, becomes
more and more critical in the daily business operation. Customers cannot
afford to have their Domino servers fail for hours. This may affect customer
service levels.
Lotus Domino for AS/400 has already taken advantage of the reliability and
availability features of the AS/400 system, such as RAID-5, mirrored disk
units, and integrated backup capability. With the AS/400 system’s logical
partitioning and Domino clustering, the level of availability is further
enhanced.
Logical Partitioning Concepts
9
Partition 1
SYS A
SYS B
Domino Server 1
Primary
Partition
Partition 2
SYS C
Domino Server 2
lotus Notes
Database Replica
Lotus Notes
Database
Domino Clustering via Virtual
OptiConnect
Cluster-aware
Notes Clients
Figure 7. Lotus Domino Clustering over AS/400 Logical Partitions
The cluster-aware Notes clients in Figure 7 use Domino server 1 as the home
server. When the Notes client tries to open a database on the Domino server
1 and the server is down, the clients look for the next available server on the
server list in the client’s own cluster cache. Then, they access the Cluster
Manager on Domino server 2 in the cluster. The cluster manager then
redirects the client to Domino server 2. The user continues to have the
database access on server 2. The cluster replication ensures that all the
changes to the database are immediately passed to the database replica in
the cluster. Databases are tightly synchronized to provide the high availability.
Putting the Domino servers into two different AS/400 logical partitions
isolates them from software errors affecting both servers. Plus, it also
provides the ability to have high-speed cluster replication between the two
AS/400 partitions using the virtual OptiConnect function.
1.2.9 Three-Tiered Application Architecture
Several types of AS/400 applications use a three-tiered architecture. For
example. Enterprise resource planning (ERP) applications often use a
desktop PC, an application server and a back-end database server. With the
availability of AS/400 logical partitioning, customers can put the database
server in the primary partition and application servers in secondary partitions
within the same system. This can reduce footprint and maintenance costs.
10
Slicing the AS/400 with Logical Partitioning: A How to Guide
Furthermore, the database server can communicate with the application
servers through the virtual OptiConnect function available between the
partitions.
1.3 How It Works
Logical partitioning (Figure 8) is achieved by distributing the resources of a
single AS/400 system to create multiple independent systems, within the
same system. The resulting structure consists of a primary partition, and one
or more secondary partitions.
SYS A
OS/400
PPPP
MMMM
MM
Logical
Partitioning
SYS A
OS/400
SYS B
OS/400
SYS C
OS/400
P
MM
PP
MMM
P
M
Primary
Partition
Secondary
Partition 1
Secondary
Partition 2
Primary
Partition
P = Processor, M = Memory
Figure 8. AS/400 Logical Partitioning
Resources, such as processors, memory, and I/O devices, are logically
assigned to primary and secondary partitions.
1.4 Partitioning Terminology
The new nature of logical partitioning demands that all terms in use be clearly
understood. Let us review the hardware and software components that we
will use to create and manage logical partitioning, and define some new
terms.
This book is structured to offer you the flexibility of browsing. At this point,
you may want to refer to Appendix A, “Logical Partitioning Terminology” on
page 167. This appendix includes the terms that we are going to use in the
process of working with logical partitioning. You may also want to refer to
Figure 9 on page 13 to see where everything fits into the context.
Logical Partitioning Concepts
11
1.5 Primary Partition
When OS/400 is installed on a new system, it is always configured with the
primary partition only. This partition owns all the resources available on the
machine, such as processors, main storage, and system buses. It functions
as one of the logical systems and also provides management functions for
the configuration of the secondary partitions, such as:
• Power Management (includes concurrent maintenance)
• Virtual Operations Panel (controls power up and down, IPL, and virtual key
lock)
• Logical Partition definition (for configuring logical partitions)
The primary partition can also function as a hub for external communications
or Electronic Customer Support for the entire system.
1.6 Secondary Partitions
Secondary partitions are created and managed from the primary partition, but
function as independent systems within the physical system. They have their
own resources such as processors, main storage, and system buses. Plus,
they have their own primary language, system values, time-of-day, user
profiles, OS/400 system Licensed Internal Code (SLIC), IBM Licensed
Programs (LPP), database files, and user applications. Furthermore, they can
be independently powered down and up without affecting the primary or other
secondary partitions. However, they retain some dependencies on the
primary partition. For example, performing a system restart or power down on
the primary partition powers down all secondary partitions.
This means that, before such actions are undertaken, all secondary partitions
must be powered down to avoid possible abnormal secondary partitions
termination.
1.7 Conceptual View of Logical Partitioning
Once you have reviewed the most common terms related to logical
partitioning, we can see how it all fits together in Figure 9.
12
Slicing the AS/400 with Logical Partitioning: A How to Guide
Prim ary
SLIC X
H yp
e rv
i so
Secondary
Secondary
SLIC Y
SLIC Z
r
Virtual OptiConnect
PLIC
B us
Tape
IB
CD
M FIOP x
IOP y
IOP z
M
Loa d
Source
LAN
Load
Source
IOP x
IOP y
Loa d
Source
LAN
IOP z
Tape
I BM
LAN
CD
IOP xyz
Figure 9. Generic View of Logical Partitioning
Figure 9 shows how a single system can be configured to have three
partitions, with one as the primary partition. The logical partitions are also
interconnected using the AS/400 system bus architecture through virtual
OptiConnect to achieve high-speed data transfer between the partitioned
systems. Each partition shows some dedicated I/O resources, such as disk
drives, communications links, and workstation controllers. The example also
shows additional resources, such as tape drives, CD-ROM drives, and
communications adapters that can be dynamically shared by all of the
partitions.
1.8 Clustering
Logical partitioning creates independent systems within a single physical box.
Clustering can be seen as a superset of logical partitioning in that it provides
a single resource view that binds together two or more physical AS/400
systems. These can, in turn, be logically partitioned if required.
Logical Partitioning Concepts
13
Clustering offers increased high availability of applications and data. It does
this by providing automated recovery facilities across several combined
AS/400 systems (defined as nodes in the cluster). For example, through
clustering, an application can be set up to run on a primary node, with a
secondary node defined for quick fail-over switching, when the primary node
becomes unavailable. The switch-over would include the automatic switching
of communications to the secondary node.
It is also possible to imagine a cluster containing logically partitioned nodes.
An application can run in one partition, in its own primary node, and have its
backup secondary node ready on a logical partition in another node,
somewhere in the cluster.
Figure 10 illustrates the concepts of clustering.
5-Node Cluster
Partitioned
Node
Cluster Control
LAN
Figure 10. Clustered Configuration with a Partitioned Node
14
Slicing the AS/400 with Logical Partitioning: A How to Guide
1.9 Marketplace Positioning
As we mentioned at the beginning of this chapter, system partitioning has
now become a must in the server marketplace. It is interesting to look at
some of the active players that offer partitioning solutions (see Table 1).
Table 1. Logical Partitioning Marketplace Offerings as of February 1999
S/390
AS/400e
Operating Systems
All
OS/400
All Enterprise
6xx, Sxx, 7xx
Servers
Partitioning Type
Logical
Logical
Maximum # partitions
15
12
Maximum # processors / partition
12
12
Minimum # processors / partition
<1
1
Processor Increment
<1
1
Supported Servers
above minimum
Memory Increment
above minimum
I/O Increment
above minimum
Independent Movement of
Resources
Dynamically Change Resources
Partition Weights
High speed internal
communications
SUN*
HP 9000*
Solaris
HP-UX
E10000
None
Physical
8
64
1-4
None
1-4
1-4 MB
1MB
512 MB
Device
IOP
2 Buses
Yes
Yes
No
Yes for MVS
Planned,
Stage 2
Yes
database restart
required to use
Yes
Interactive
No
No
Yes
No
1.10 What Makes the Difference
In 1.2, "Scenarios for Logical Partitioning" on page 2, we discuss some of the
ways in which logical partitioning can be beneficial. The reality is that the
permutations are limited only by your imagination and circumstances.
The AS/400 system now offers many more options that can be manipulated to
meet every possible customer requirement involving physical machine
consolidation and logical machine independence. The entire offering is
packaged using the best technology available in the market place. To top it
all, the AS/400 system product line has a wide and open-ended range that is
capable of satisfying from the simplest to the most complex customer
environments imaginable. Keep reading the remaining chapters of this book
to discover how you can put it all together.
Logical Partitioning Concepts
15
16
Slicing the AS/400 with Logical Partitioning: A How to Guide
Chapter 2. Planning Considerations
Detailed planning for logical partitions is crucial. The effort made during the
planning stages makes the transition to a partitioned environment a smoother
process. Insufficient planning results in problems during the configuration of
the new environment.
This chapter summarizes the areas you need to consider when planning for
logical partitions. It is not intended as a replacement to the formal planning
process that is documented on the logical partitioning home page at:
http://www.as400.ibm.com/lpar
This Web site contains the latest information available and provides links to
these three essential articles:
• Planning for and Setting Up Logical Partitions
• Backing Up and Recovering Logical Partitions
• Managing Logical Partitions
2.1 Understanding Customer Requirements
Logical partitioning is a new mode of operation for the AS/400 system where
multiple copies of OS/400 run on a single physical set of hardware.
Physically, the system will be packaged in exactly the same way that it is
today. Multiple racks contain various types of I/O devices. However, each
system has only one main system rack that contains processors, memory,
and the system control panel.
Logical partitions can be considered a suitable solution for many different
requirements. In some circumstances, partitions only meet some of the
customer’s overall requirements. The requirements that are not met need to
be identified. The impact of this to the business should be evaluated.
2.1.1 Single Points of Failure
A machine that is configured for logical partitions has points of failure that can
affect all partitions on that physical machine. A processor or memory card
failure in any partition can cause the entire machine to end abnormally. Then,
each partition has to perform an IPL to recover.
Any software or hardware problems that cause the primary partition to fail
also cause all the secondary partitions to fail.
© Copyright IBM Corp. 1999
17
2.1.2 Availability Requirements
When using logical partitions for AS/400 hardware or application
consolidation, the planning must balance the optimum consolidation possible
with the customer’s availability requirements.
It is possible to have a scenario where data from one partition is replicated to
another partition on a different AS/400 system. This can be done using one of
these options:
• High Availability Business Partner (HABP) applications
• The remote journalling function of OS/400 with a user defined process to
apply the journal entries on the target partition
• A customer-written application
All three approaches can use either traditional LAN/WAN communications or
the Opticonnect/400 function discussed in Chapter 7, “Inter-Partition
Communication” on page 133.
Each AS/400 system then acts as a partial standby machine to the other. In
the case of a software failure or specific hardware failure, only the workload
from that partition has to be moved to the backup partition on the second
physical machine.
18
Slicing the AS/400 with Logical Partitioning: A How to Guide
S ys tem A
P a rtitio n s
0
1
2
S o lu tion s
H A B P A p p lica tion s
R e m o te J o u rn a llin g
C u sto m e r w ritten
U s in g
O ptiC o nn e c t/4 0 0
L A N /W A N
S ys tem B
0
1
2
P a rtitio n s
Figure 11. Replicating Data from One Partition to Another
Applications that have vastly different availability requirements may not work
well together on the same AS/400 system.
2.1.3 Resource Requirements
Applications that have opposing peak resource times would fit well together
on the same machine. Consider the example where application A has a high
CPU requirement during the online day but very little overnight. This can be
consolidated with an application that has very little CPU requirement during
the day but a high requirement overnight. Appropriate planning is required to
factor in the time needed to move CPU and memory resources between
partitions.
2.1.4 Application Stability
The stability of a customer’s applications may impact the selection of
applications that run in separate partitions on the same AS/400 system.
An application that is fairly new into production may have yet to reach an
adequate level of stability. This instability may impact the operational
requirements and level of day-to-day management required.
Planning Considerations
19
Applications that have similar upgrade and maintenance requirements are
better suited on the same AS/400 system.
2.1.5 Security Requirements
Every application and every system has its own security requirements.
Appendix C, “Accessing LPAR Functions” on page 175, provides a list of all
partioning functions and indicates from where they can be accessed.
Consider this information when selecting applications to run in different
partitions on the same AS/400 system.
2.2 Understanding the New Complexity of Your Environment
It is important to understand the new operating environment that is created as
a result of setting up partitions. New functions, such as the fast inter partition
communication discussed in Chapter 7, “Inter-Partition Communication” on
page 133, open up more options to developers and support staff alike. These
new functions, plus the inter partition dependencies for IPL and maintenance
lead to new complexities for operational staff. The following sections describe
some of the complexities of each management discipline.
Operations Management
Operators are introduced to new terminology and possibly new tasks to
perform from unfamiliar menus. Technical support staff must be aware of the
new environment and the fact that a change in one partition may impact
another partition. For example, with the current implementation, allocating
more memory to one partition requires it first to be removed from another.
Problem Management
Help desk staff must be aware of the new environment so they can perform
comprehensive problem determination. The performance for inter partition
communication is far better than what is available today using external
LAN/WAN communications. Applications that read and update data between
partitions are likely to use these new facilities.
Change Management
Application changes may need to be implemented a different way. See 4.2,
“Managing User Applications in a Logical Partition” on page 63, for more
details.
Performance Management
Performance and capacity planning need to take into account partition
processing as well as application performance. See Chapter 6, “Capacity
Planning and Performance” on page 81, for more details.
20
Slicing the AS/400 with Logical Partitioning: A How to Guide
Configuration Management
Any system configuration changes that are made are likely to impact all
partitions. Extra planning is required when trying to implement a change to
the system. The dynamics of managing multiple systems have not changed
with logical partitioning. You still need to view the system management
disciplines across a logically partitioned system in exactly the same manner
as you would do with multiple AS/400 systems.
2.3 Logical Partitioning Configuration and Ordering Process
The following steps outline the process to be followed when IBM or a
business partner wishes to configure and implement logical partitions for a
customer. The actual process may vary from customer to customer.
1. Find out more on logical partitioning. For further information, refer to the
logical partition home page at: http://www.as400.ibm.com/lpar
2. Go to the logical partition home page, and print and review the three
essential articles:
• Planning for and Setting Up Logical Partitions
• Backing Up and Recovering Logical Partitions
• Managing Logical Partitions
3. Go to the logical partition home page (see step 1), and review the AS/400
Logical Partitions Hardware Planning Guide.
4. Perform a capacity plan from a logical partitioning perspective.
The normal capacity planning process needs to take into account the
performance available from the resources allocated to a partition. For
more information, refer to Chapter 6, “Capacity Planning and
Performance” on page 81.
5. Complete the Logical Partition Planning Worksheet.
The planning worksheet is a framework for providing the information
necessary to build an AS/400 system capable of supporting partitions.
This can be found as a link from the logical partition home page and
should be printed and then completed.
6. If partitioning is to be implemented on an existing AS/400 system, print an
accurate hardware rack configuration. Follow these steps:
a.
b.
c.
d.
Enter the STRSST command.
Select option 1 (Start a Service Tool).
Select option 7 (Hardware Service Manager).
Press F6 (Print configuration).
Planning Considerations
21
If partitioning is to be implemented on a new AS/400 system, print the
configuration report from the PC configurator.
7. Contact the Rochester Technology Solutions Center (TSC) and supply
them with the completed paperwork from step 4 and 5.
They can be contacted by e-mail at: [email protected] or by fax at:
1-507-253-0424.
The IBM Representative or Business Partner will arrange for a technical
conference call between all parties if this is required.
8. Conduct a Systems Assurance review with the IBM Representative,
Customer and Business Partner.
9. Order the necessary hardware and software (V4R4M0).
The TSC will return a list of additional hardware that needs to be ordered
to support the planned partitioning environment.
10.Install V4R4M0.
We recommend that V4R4M0 is installed and run for one complete
business cycle before any other changes are made to the system. This
allows V4R4M0 to stabilize in the customer environment.
11.Install the new hardware.
The TSC will provide the necessary card placement information necessary
to support the specific partition requirements. This may require moving
some existing hardware from their current card positions.
12.Set up the logical partitions.
Do this by following the process described in using the Planning for and
Setting Up Logical Partitions article.
13.Verify that all the hardware resources are reporting correctly in each
partition.
Refer to 3.11, “Managing the Hardware Resource Configuration” on page
53, for more details on how to do this.
14.Print the system configuration and store it with the service and disaster
recovery procedures.
Refer to 3.11.4, “Printing the System Configuration” on page 58, for details
on how to do this.
22
Slicing the AS/400 with Logical Partitioning: A How to Guide
2.4 Hardware Considerations
Every logical partition requires mandatory hardware to function properly.
Logical partitions can also be configured with other optional hardware, some
of which can be switched between partitions. Before planning the detailed
hardware requirements for a system, it is important to understand the basic
hardware requirements.
Partitioning can only be implemented on an Nway AS/400 processor. Each
partition requires:
•
•
•
•
•
A minimum of one processor
A minimum of 64 MB of memory (256 MB in the primary partition)
A console device for each partition
A load source DASD
An alternate IPL device (this may be a device that can be shared with
other partitions)
• IOPs that support the above three devices
Full details of the hardware supported with partitioning can be found in the
Logical Partitions Hardware Planning Guide accessible from the logical
partitioning home page at: http://www.as400.ibm.com/lpar
2.4.1 Supported AS/400e Models
Partitioning is only supported on the AS/400e series hardware. This
restriction is due to processor architecture on earlier AS/400 models.
Table 2. Supported Models
AS/400 Model Number
Processor Feature Codes
AS/400e server S20
#2165, #2166
AS/400e Custom Mixed-Mode server S20
#2170, #2177, #2178
AS/400e server S30
#2258, #2259, #2260
AS/400e Custom Mixed-Mode server S30
#2320, #2321, #2322
AS/400e server S40
#2256, #2261, #2207, #2208
AS/400e Custom Mixed-Mode server S40
#2340, #2341
AS/400e model 620
#2182
AS/400e model 640
#2238, #2239
AS/400e model 650
#2240, #2243, #2188, #2189
Planning Considerations
23
AS/400 Model Number
Processor Feature Codes
AS/400e server 720
#2063, #2064
AS/400e server 730
#2066, #2067, #2068
AS/400e server 740
#2069, #2070
Partitioning offers IOP-level partitioning, as well as bus-level partitioning. The
IOPs used on a shared BUS must be able to IPL independently. For a current
list of supported IOPs, refer to the Logical Partitions Hardware Planning
Guide (available from the logical partition home page).
During hardware planning, and with the movement and placement of IOP, you
may need to purchase additional storage expansion towers or system
expansion towers. Each additional tower is considered a separate bus.
Ensure that the total number of buses across the entire system is not
exceeded.
2.5 Software Considerations
Logical partitioning is supported in Version 4 Release 4 Modification 0
(V4R4M0) and subsequent versions of the OS/400 operating system. It is not
supported on earlier releases of OS/400. Any systems that implement logical
partitions in this first stage will run all partitions at V4R4M0 of OS/400.
2.5.1 Licensed Program Products — Licensing
A customer requires only one license of OS/400 per physical system,
regardless of the number of partitions created. For licensed program products
(LPP) that are licensed by processor group, the customer purchases one
license according to the processor group to which the physical system
belongs, regardless of the number of partitions created. The LPP can then be
installed in each partition on the system if required.
For LPPs that are licensed based on the number of users, the customer
purchases a license according to the number of users required across the
entire AS/400 system. The LPP can then be installed in each partition on the
system if required.
In V4R4M0, license management for user-based LPP is checked only at the
individual partition level. In future releases, this will be done at the system
level, meaning the total of all users of the LPP in all partitions.
24
Slicing the AS/400 with Logical Partitioning: A How to Guide
The following LPPs have user-based pricing:
• 5769-CL3
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
5769-CP4
5716-CX4
5769-CX5
5769-DFH
5769-FW1
5769-RD1
5769-SM1
5769-SS1
5769-SVD
5769-SVE
5769-SV3
5798-TBT
5769-VI1
5769-VI1
5769-WP1
5769-XW1
5769-XY1
5769-XZ1
VisualAge RPG and Cooperative Development Environment
for AS/400
Call Path Server for AS/400
VisualAge C++ OS/2 Client
(Option 1) VisualAge C++ Windows 95/NT Client
CICS for AS/400
Firewall for AS/400
(Option 5) OnDemand Server Feature
System Manager for AS/400
(Option 25) OS/400 NetWare Enhanced Integration
Netfinity AS/400 Manager for OS/2
Netfinity AS/400 Manager for Windows 95
ADSTAR Distributed Storage Manager for AS/400
OfficeVision JustMail
ImagePlus VisualInfo for AS/400
(Option 2) ImagePlus VisualInfo Object Server
OfficeVision for AS/400
AS/400 Client Access Family For Windows
AS/400 Client Access Family
OS/2 Warp Server for AS/400
Third-party products are licensed according to the policy of the individual
vendors.
2.5.2 Software Problems That Can Affect All Partitions
Logical partitioning documentation may mention some software failures
affecting the entire machine or other partitions. This refers to a software
failure that is occurring in the Partition Licensed Internal Code (PLIC). The
frequency of this happening is regarded as very low due to the thorough
testing of all releases.
2.5.3 Installing V4R4M0
We recommend that V4R4M0 is installed and allowed to stabilize in the
customer’s environment before logical partitions are implemented. As with
any upgrade to OS/400, the customer should run for at least one business
cycle. This will ensure that, in the unlikely event any problems are
encountered, they can be resolved before any further changes are made to
the system.
Planning Considerations
25
2.5.4 Application Dependencies between Partitions
There are no application dependencies when using partitions. Each partition
can operate totally independently of the others. If required, they can have
inter-partition communication between selected partitions. There is nothing in
the partitioning function that changes the way that application code runs on
the system.
2.5.5 National Language Considerations
There are no special national language considerations for logical partitioning.
Secondary languages are installed in the same way as they are today.
The situation may arise where a customer requires a different primary
language for each partition created on the system. This is allowed under the
terms of the licensing agreement. As the time this redbook was published, it
was not known how this will be supported by customer fulfillment. For further
information, please contact your local IBM representative.
Some countries choose to translate the dedicated service tool (DST) displays
into the relevant national language. This may influence the primary language
that is installed on a partition. The majority of logical partitioning configuration
and maintenance tasks are performed from the primary partition dedicated
service tool displays. However, some service-related tasks can only be
performed from dedicated service tools on the secondary partitions.
2.5.6 Future Release Co-Existence
When planning for future release co-existence on the same physical
hardware, consider the minimum level of OS/400 required to support any new
hardware that becomes available.
2.5.7 Published Software Limits
The OS/400 software maximums are the limits per partition. For example,
each partition can support up to a maximum of 16 auxiliary storage pools
(ASPs).
2.6 Interface to Create and Manage Partitions
You can only restart, reconfigure, and manage logical partitions from the DST
and system service tools (SST) environment. For information on which
options can be accessed from each environment, please refer to Appendix C,
“Accessing LPAR Functions” on page 175.
26
Slicing the AS/400 with Logical Partitioning: A How to Guide
DSTs are a set of tools normally used to work with the system when the
operating system is not available. SST are a subset of the DST tools. The
tools available through SST, such as displaying the system configuration, can
be used while the operating system is running and other users are on the
system.
These environments do not provide the same user interface and support that
is provided with standard OS/400 displays. For example, there is no
cursor-sensitive help, second-level information available for messages, or
prompting of options or commands. There are many system maintenance
functions that can be performed from these menus that assume the user is
aware of the impact that their actions will cause.
People who are not familiar with using this interface find the management of a
partitioned environment very different from what they are used to using. The
single, most important point to remember when using the DST or SST
interface is the message that is presented when you first select the "Work
with system partitions" option from the main DST or SST menu.
Attention
Incorrect use of this utility can cause damage to data in this system. See
the service documentation that comes with the system. For more
information on how to use DST and SST, access the IBM AS/400 Technical
Studio at: http://www.as400.ibm.com/tstudio
If you are not familiar with the DST and SST environment, we recommend
that you do not attempt to configure LPAR on your own. Assistance from IBM
is available under the terms and conditions of a normal services contract.
The DST and SST interface is different from what the majority of customers
use. We recommend that you take this into account when planning the level
of IOP switching that is going to occur between partitions.
The parameter settings specified when setting up and creating partitions can
have a major impact on the management and performance of the partitions.
Refer to Chapter 3, “Operational Considerations” on page 39, to understand
the full implications of these parameters.
Planning Considerations
27
2.7 Deciding What Runs in Primary and Secondary Partitions
As defined in 1.5, “Primary Partition” on page 12, a primary partition exists on
every N-way AS/400e model. It is possible to create multiple secondary
partitions depending on the resources available.
The primary partition is a fully functional system with additional responsibility
to manage all secondary partitions. All partition management functions are
included in the Partition Licensed Internal Code (PLIC) of the primary
partition.
Secondary partitions are logical systems created from the hardware
resources that are available to them. Secondary partitions are independent of
each other, but maintain a dependency on the primary.
Due to the significance of the primary partition, take care when deciding what
workload to run in that partition. You have three choices of the type of work
that can run in the primary partition:
• Nothing — The primary partition is configured with the minimum amount
of resources allowable and purely performs partition management tasks.
• Production — The primary partition is configured to run a production
environment. The partition resources are allocated to allow for the system
to run the selected workload, plus partition management tasks.
• Test — The primary partition is configured to run a test environment. The
partition resources are allocated to allow for the system to run the selected
test workload, plus the partition management tasks. This is not
recommended due to the instability that normally exists in this kind of
environment.
There is much discussion that can be had about the pros and cons of each
choice that cannot be adequately covered here. The decision of what
workload is run in each partition should take into account these aspects:
• Stability of the application — An application that has not reached
stability or is going through frequent change can impact the partition in
which it runs. Performance can be erratic if programs go into loops, or if
errors are constantly being encountered that have to be handled by the
operating system.
• Availability requirements — Customers who have a high availability
requirement on any of the secondary partitions should not run applications
in the primary partition. High availability planning requires that all possible
steps are taken to minimize the impact any problems encountered can
have on the applications requiring high availability. An application problem
28
Slicing the AS/400 with Logical Partitioning: A How to Guide
encountered in the primary partition can impact the partitions
performance, operating system, or LIC. Such a problem can also cause a
system failure. A failure in the primary partition can, therefore, affect the
secondary partition and the applications that are running in the secondary
partition.
• Application testing — An error in an application function does not
normally result in the failure of an AS/400 system. Errors encountered
normally result in OS/400 exception messages being issued and the
function ending abnormally. If the problem is found to be in IBM generated
code, then a PTF is required to fix the problem. Testing applications that
use leading edge functions that are likely to require frequent PTFs should
not be done in the primary partition. The disruptive nature of installing
PTFs can affect all secondary partitions. This is particularly true when
PTFs require a system IPL on the primary partition.
• IBM code testing — New release testing or PTF testing should not be
done in the primary partition if production applications exist in the
secondary partitions. In very rare circumstances, defective PTFs can
effect all secondary partitions. Using the primary partition for this sort of
testing is similar to installing new releases of OS/400 or cumulative PTF
packages directly onto a production machine.
• Allocation of resources — Allocation of resources can be performed
from within the DST or SST environment. Incorrect use of the utilities on
these menus can cause damage to data on the system. Consider the skill
level of the personnel likely to require access to the DST and SST
environment. Applications should not be run in the primary partition if it
requires personnel who are unfamiliar with the DST or SST environment to
perform maintenance through these screens.
• Security of partitions — You can restart, reconfigure, and manage all
logical partitions from the primary partition by using DST or SST. If
applications require personnel to have access to the DST or SST function,
remember that they can affect all logical partitions. We strongly
recommend that you change and secure DST and SST passwords.
2.8 Choosing Bus- or IOP-Level Partitioning
Depending on the requirements, there can be advantages to setting up one
type of logical partitioning over another. There are two types of partitions that
can be created:
• Bus-level partitions
• IOP-level partitions
Planning Considerations
29
For a full definition of each type, refer to the article Planning for and Setting
Up Logical Partitions article. You can find it on the logical partition home page
at: http://www.as400.ibm.com/lpar
Consider the points described in the following sections when deciding
whether to have Bus or IOP level partitions.
2.8.1 Bus-Level Partitions
Bus-level partitions offer the following benefits:
• Better problem isolation between partitions since each bus and all the
devices attached to that bus do not depend on other partitions
• Better performance since SLIC does not have to call PLIC to queue
commands for a bus
• Less day-to-day management of IOP resources since they are
permanently allocated to the partition
• Limited requirements for personnel unfamiliar with DST/SST environment
to access DST functions
Note: The total number of buses that an AS/400 system can have is 19,
regardless of how partitions are defined. You should plan accordingly.
2.8.2 IOP-Level Partitions
IOP-level partitions offer the following advantages:
• Allows resources to be switched among multiple partitions. This can help
reduce the cost of hardware required.
• IOP resources need to be managed from the SST or DST environment,
but do not require a system or partition IPL to become active. Incorrect
use of the utilities on the DST and SST menus can cause damage to data
on the system. Consideration should be given to the skill level of the
personnel likely to require access to the DST and SST environment.
• Any bus problems encountered can potentially affect multiple partitions.
Note
Regardless of the type of partioning selected, the primary partition controls
the availability of each bus on the system. If a secondary partition owns a
shared bus, and that partition is powered off, all IOPs on that shared bus
dedicated to other partitions can still be used.
30
Slicing the AS/400 with Logical Partitioning: A How to Guide
2.9 Moving to a Partitioned Environment
A customer wanting to move from an existing environment to a partitioned
environment needs to plan to minimize the impact to the existing users. You
can do this by choosing one of the following options:
• Install a new system.
The simplest approach is to install a new AS/400 system and set up the
partitions ahead of time. The workload that is going to run in each partition
can then be migrated at a convenient time to the new system. As each one
of the existing systems is moved to the new system, the hardware from the
existing system can be added to the new system ready for other partitions
to use.
• Replace the release on the current system.
The existing system is upgraded to V4R4M0 of OS/400 and the existing
workload runs in the primary partition. After the stability of V4R4M0 is
proven, secondary partitions can be created and workloads migrated to
these new partitions. Careful planning is required since you may have to
remove configured disks so that they can be assigned to the secondary
partition.
• Upgrade the current system.
The existing system is upgraded to V4R4M0 of OS/400 and the existing
workload runs in the primary partition. After the stability of V4R4M0 is
proven, additional hardware is added to the system. This hardware is used
when creating secondary partitions. The workload can then be migrated to
the new partitions.
2.10 System Naming Convention
The default system naming convention works the same way as it does today.
The primary partition has a system name of Sxxxxxxx, where xxxxxxx is the
seven-character system serial number that can be displayed using the
following command:
DSPSYSVAL QSRLNBR
Since each system name generated must uniquely identify each partition
within the physical system, all secondary partitions created will generate
default system names in the form of Axxxxxxx through Kxxxxxxx.
The system name can be changed using the Change Network Attribute
(CHGNETA) command. It does not depend on the partition name specified
Planning Considerations
31
when creating or changing a partition. However, to simplify matters, we
recommend that you keep the partition name the same as the system name.
A change to the partition name must be done from DST on the primary
partition, but does not require an IPL.
2.11 System Serial Number
The system serial number is the same across all partitions. This is commonly
used as a unique system identifier. The system serial number has to be
qualified by the partition ID to achieve the equivalent effect.
The primary partition always has a partition ID of zero. Secondary partitions
then start from 01 and increment according to the number that are created.
In Appendix B, “Sample RPG and C Programs” on page 171, there are two
sample programs that show how to retrieve the individual partition
information.
2.12 Security
You can only restart, reconfigure, and manage logical partitions from the DST
and SST environment.
Systems that are set up to run partitions may require more access to the DST
and SST environments than is typically necessary. As explained earlier in this
chapter, these environments provide powerful tools that can cause major
damage to your data if used incorrectly. We recommend that the default
passwords for the DST and SST environment be changed to prevent
unauthorized access.
We also recommend that the key is not left in the operators control panel.
Otherwise, this would allow any unauthorized person to force the DST signon
screen to the console of the primary partition.
Using non-partition related functions in DST and SST is restricted purely to
the resources allocated to that partition. For example, it may not be possible
for one partition to display or alter the contents of another partition’s memory
or disk.
32
Slicing the AS/400 with Logical Partitioning: A How to Guide
2.13 Updating Your Save Strategy
As you partition a system, you have to consider the requirements of all
partitions when you develop your save and restore strategy. You should
consider the following points when developing your strategy.
Consider Using a Tape Library
Having a physical system with multiple partitions and possibly switchable tape
resources between these partitions is a more complex environment to
manage from the save and restore perspective. We recommend that you
consider a tape library and BRMS/400 as a way to help manage the new
environment.
Availability of the Tape Drives
Different applications have different save and restore requirements. Because
a system has a certain number of tape drives attached to it, does not mean
that all tape drives are available for all partitions to use at anytime. If you
switch removable media devices between partitions, then your backup of
these partitions will become a sequential process.
Accessing More Hardware
It may be possible to have more hardware available for you to use on your
save or restores.
The top half of Figure 12 on page 34 shows that System A, System B, and
System C all have connections to the four 3590 model B11 tape devices. You
can connect a tape unit 3590 Model B11 to a maximum of two AS/400
systems using the dual port on the tape control unit. The only system that can
access all four tape drives is system B.
If these three systems were consolidated onto a single AS/400 system but
with three partitions, as shown in the lower half of Figure 12, it would be
possible for all tape drives to be available to each partition. Then, the tape
drives have to be connected to the system using shared buses so that all
partitions could use them.
Continue to follow any configuration recommendations for performance to
ensure the same performance is achieved on a partitioned system as it would
be on a traditional system. For further information, refer to the AS/400
Performance Capabilities Reference Guide. You can request a copy of this
document by downloading it from the AS/400 On Line Library through the
AS/400 Internet site at: http://www.as400.ibm.com
Planning Considerations
33
The document is viewable and downloadable in the Adobe Acrobat (.pdf) format
and is approximately 1.5MB in size. The Adobe Acrobate Reader plug-in is
available at: http://www.adobe.com
System A
System B
System C
Partitions
0
1
2
Figure 12. Using Multiple Tape Drives
Taking Saves from a Second Copy of the Data
Using one of the High Availability Business Partner (HABP) applications, it is
possible to replicate data from one partition to another on the same machine.
Although this option does not offer any higher level of availability with
unplanned outages, it offers a higher level of availability with planned outages
such as daily backup. The production machine can always be available to the
users, while the second copy of data is used for the backup to tape.
The CPU required for this partition will have to be adequate so that data
changes from other partitions can be received and applied and when
required, the save tasks can be run. A save task is typically I/O intensive so
this in itself will require very little CPU cycles.
Enough DASD needs to be installed in this partition to support the amount of
data it will be required to store. This is likely to be the largest partition on the
system.
34
Slicing the AS/400 with Logical Partitioning: A How to Guide
When a save is required, the apply process in the backup partition is stopped
and the save taken from this copy of the data. Work continues to run in the
original partition and journal entries generated by changes to the data are still
sent to the backup partition but not applied.
Partition0
Partition1
Partition2
6processors
4processors
2processors
High AvailabilityBusinessPartner Software
24x 7availabilityfor planned outages
Used for savesand as astandby
Figure 13. Saving from a Second Copy of the Data
Using ObjectConnect
ObjectConnect/400 saves information from one system and restores it to
another using a single command. ObjectConnect is a non chargeable product
option of OS/400. For more information on ObjectConnect/400, refer to 7.2.2,
“ObjectConnect/400” on page 137.
ObjectConnect can be used to snapshot data at a particular point in time. The
users stop using the application so a copy of the data can be transferred to
another partition. Later, this data can be saved to removable media.
Planning Considerations
35
Partition 0 Partition 1
Partition 2
6 processors
2 processors
4 processors
SAVRSTLIB
SAVRSTOBJ
SAVRSTCFG
SAVRSTCHG
SAVRSTDLO
SAVRST
Figure 14. Using ObjectConnect to Move Data to Another Partition
The following test results demonstrate the gigabyte-per-hour (GB/HR)
transfer rates that can be achieved when transferring data between partitions
using ObjectConnect/400 and the virtual bus adapter. OptiConnect/400 or the
OptiMover PRPQ (5799-FWQ) must also be installed to enable
ObjectConnect to use the virtual high-speed bus adapter. Inter partition
communication must be enabled on both partitions before this can be used.
For more information, refer to Chapter 7, “Inter-Partition Communication” on
page 133.
The tests were performed on a model 720 #2064 divided into two partitions.
Each partition had two processors and 4096 MB of main storage.
Test 1 used one SAVRSTLIB command to save a single library from one
partition and restore it to another. The library contained 22 objects and was
2.7 GB in size. This job transferred the data at the equivalent rate of
approximately 27 GB/HR.
Test 2 used one SAVRSTLIB command to save 20 libraries from one partition
and restore them to another. Each library contained 22 objects and the total
amount of data transferred was 54 GB. This job transferred the data at the
equivalent rate of approximately 28 GB/HR.
36
Slicing the AS/400 with Logical Partitioning: A How to Guide
Test 3 used twenty concurrent SAVRSTLIB commands to each save a single
library from one partition and restore it to another. Each library contained 22
objects. The total amount of data transferred was 54 GB. These jobs
transferred the data at the equivalent rate of approximately 28 GB/HR
As the results show, it is possible to consistently achieve a save and restore
rate similar to what a 3590 tape drive can achieve.
2.14 Published Hardware Limits
The AS/400e hardware maximums published in the AS/400e System
Handbook, GA19-5486, are the limits per physical system. The sum of
individual hardware types across all partitions must not exceed the published
limit for a physical system.
2.15 DASD Compression Support
DASD compression support is the same on a partitioned system as it is on a
non-partitioned system. Compressed DASD is only supported in user ASPs.
DASD compression must be turned off for a disk or disk IOP before it is
removed from a user ASP, if the DASD is going to be used for the system
ASP on another partition.
2.16 Dual Tape Attach Support to Multiple Partitions
It is possible to attach some model tape devices to two AS/400 systems. Only
one system at a time can use the device.
On a partitioned system, it is possible to attach these drives to two partitions
using separate IOPs. In the event that these IOPs are subsequently
dedicated to the same partition, the same usage rules apply.
The system only allows one device description to be varied on at any one
time. This allows the second device description that uses the second IOP to
remain inactive.
Planning Considerations
37
38
Slicing the AS/400 with Logical Partitioning: A How to Guide
Chapter 3. Operational Considerations
This chapter offers useful information to help you get ready for and run a
system with multiple partitions.
3.1 Installing New Hardware Ready for Partitioning
The Technology Solutions Center (TSC) provides a list of any new hardware
that is required to support the planned partitioning environment. This
hardware needs to be installed before the partitions can be created.
The IOP placement information is also provided by the TSC. The hardware
placement information should be reviewed with the customer engineer (CE)
prior to installing the new hardware. This ensures that it is installed in the
correct place according to the specific partition requirements.
The following sections explain how the hardware is treated by the system
when it is installed. This assumes V4R4M0 has been installed.
Processors and Main Storage
If processors or main storage are added to a system that has no other
partitions created, other than the default of the primary, the new hardware will
be added to the primary partition. If processors or main storage are added to
a system with secondary partitions, the new hardware needs to be allocated
to a partition before it can be used.
Interactive Performance
If the customer adds interactive performance (model 7xx) to a system, the
interactive performance is available immediately. Interactive performance is
assigned by allocating a percentage of the total to each partition. For a
system with no partitions created, 100% is allocated to the default primary
partition.
The percentage amount allocated to each partition should be recalculated
after the upgrade. This ensures that the amount allocated is still correct in
light of the larger interactive performance installed.
Bus
Buses added to the system are added to the primary partition with a status of
"own bus shared". For more information on this term, refer to A.13,
“Ownership and Sharing” on page 169.
© Copyright IBM Corp. 1999
39
Important
Before installing the hardware, it is important to review the output from the
capacity plan. Any specific recommendations can be implemented during
the allocation of hardware resources when creating the partitions.
IOP
If IOPs are added to a system with multiple partitions defined. The new
hardware will be dedicated to the partition that owns the BUS.
3.2 Temporarily Making New Hardware Unavailable
If the new hardware is to be added to an existing system, it is possible that
the customer will not want the new hardware to be used until the partitioned
environment is established. Depending on the type of resource, the available
options are different.
3.2.1 Processors
To make processors temporarily unavailable, use the techniques described
here:
On the Change Partition Processing Resources display, it is possible to
change the number of processors available to a partition. If after installing the
new processors, the number available to the primary partition is reduced.
This places the additional processors into a ready, but not allocated, state.
When the secondary partitions are created later, these processors can then
be assigned to that partition.
The advantage of this approach is to ensure that correct performance
expectations are set with the users. It also ensures that they do not see a
performance penalty when the processor or any other resource is removed
from the partition in which they are working.
40
Slicing the AS/400 with Logical Partitioning: A How to Guide
Change Partition Processing Resources
System:
SYSTEMA
Type changes, press Enter.
Partition identifier . . . . . . . . . . . . . : 1
Partition name . . . . . . . . . . . . . . . . : PRIMARY
Current / available number of processors . . . : 2
New number of processors . . . . . . . . . . . . 2
Minimum / maximum number of processors . . . . . 1
/ 2
/ 2
Current / available size of main storage (MB) : 4096
New size of main storage (MB) . . . . . . . . . 4096
Minimum / maximum size of main storage (MB) . . 2048
/ 4096
/ 8192
Current / available interactive performance . : 46 / 4 %
New interactive performance . . . . . . . . . . 46 %
Minimum / maximum interactive performance . . . 25 / 46 %
Connect to system inter-partition OptiConnect . 1 1=Yes, 2=No
F3=Exit
F11=Display partition processing configuration
F9=Exclude limits
F12=Cancel
Figure 15. Change Partition Processing Resources — Processors
3.2.2 Main Storage
To make main storage temporarily unavailable, use one of the following
techniques:
• Option 1
On the Change Partition Processing Resources screen, you can change
the amount of main storage available to a partition. If after installing the
new main storage, the customer reduces the amount available to the
primary partition, the additional main storage will be placed into a ready,
but not allocated, state. When the secondary partitions are created later,
the main storage can then be allocated to that partition.
Operational Considerations
41
Change Partition Processing Resources
System:
SYSTEMA
Type changes, press Enter.
Partition identifier . . . . . . . . . . . . . : 1
Partition name . . . . . . . . . . . . . . . . : PRIMARY
Current / available number of processors . . . : 2
New number of processors . . . . . . . . . . . . 2
Minimum / maximum number of processors . . . . . 1
/ 2
/ 2
Current / available size of main storage (MB) : 4096
New size of main storage (MB) . . . . . . . . . 4096
Minimum / maximum size of main storage (MB) . . 2048
/ 4096
/ 8192
Current / available interactive performance . : 46 / 4 %
New interactive performance . . . . . . . . . . 46 %
Minimum / maximum interactive performance . . . 25 / 46 %
Connect to system inter-partition OptiConnect . 1 1=Yes, 2=No
F3=Exit
F11=Display partition processing configuration
F9=Exclude limits
F12=Cancel
Figure 16. Change Partition Processing Resources — Main Storage
• Option 2
Create and start a new subsystem description that allocates the extra
main storage when the subsystem is started. No workstation or job queue
entries should be specified to prevent the system from routing any work
through the subsystem.
3.2.3 Interactive Performance
To make interactive performance temporarily unavailable, use the following
technique:
On the Change Partition Processing Resources display, you can change the
amount of interactive performance available to a partition. If after installing
the new interactive performance the percentage available to the primary
partition is reduced, the additional performance is prevented from being used
by the primary partition. When the secondary partitions are created later, the
available interactive performance can be assigned to these partitions.
42
Slicing the AS/400 with Logical Partitioning: A How to Guide
Change Partition Processing Resources
System:
SYSTEMA
Type changes, press Enter.
Partition identifier . . . . . . . . . . . . . : 1
Partition name . . . . . . . . . . . . . . . . : LPAR2
Current / available number of processors . . . : 2
New number of processors . . . . . . . . . . . . 2
Minimum / maximum number of processors . . . . . 1
/ 0
/ 2
Current / available size of main storage (MB) : 4096
New size of main storage (MB) . . . . . . . . . 4096
Minimum / maximum size of main storage (MB) . . 2048
Current / available interactive performance . : 46
New interactive performance . . . . . . . . . . 46
Minimum / maximum interactive performance . . . 25
/ 0
/ 4096
/ 4
%
/ 46
%
%
Connect to system inter-partition OptiConnect . 1 1=Yes, 2=No
F3=Exit
F11=Display partition processing configuration
F9=Exclude limits
F12=Cancel
Figure 17. Change Partition Processing Resources — Interactive Performance
3.2.4 I/O Resources
To make I/O resources unavailable, use one of the following techniques:
• Option 1
Remove the I/O resource from the owning partition’s configuration. This
makes them available to be allocated when the secondary partitions are
created.
• Option 2
Do not create or vary off the configuration descriptions that use these
resources until they are required.
3.3 Moving Existing IOPs
If partitioning is going to be implemented on an installed system, it may be
necessary to physically move some IOPs to support the new partitioned
environment. Take care when moving IOPs to a different position on the same
system. In certain circumstances, new resource descriptions are created by
the AS/400 system. In this case, any configuration objects that are created
need to be changed to refer to the new resource description.
Operational Considerations
43
3.4 Retrieving the Resources Allocated to a Partition
In V4R4M0, there is no API to access partition resource information. There is
an unblocked Machine Instruction (MI) that can be used in a high-level
language program to retrieve this information.
The MI returns the following information about a partition:
•
•
•
•
•
•
•
•
•
•
•
•
The
The
The
The
The
The
The
The
The
The
The
The
number of partitions on the system
partition number from where this information was retrieved
logical serial number for that partition
minimum number of processors configured
maximum number of processors configured
current number of processors configured
minimum amount of memory configured
maximum amount of memory configured
current amount of memory configured
minimum percentage of interactive performance configured
maximum percentage of interactive performance configured
current percentage of interactive performance configured
Refer to Appendix B, “Sample RPG and C Programs” on page 171, for sample
RPG and C programs. These programs show how to use this unblocked MI
instruction.
3.5 Changing Partition Processing Resources
Once a partition is created, there are some changes to the configuration that,
if made, require an IPL of the affected secondary partition or an IPL or the
primary (the entire system) for the changes to come into effect. If changes are
made to the number of processors allocated, the amount of memory allocated
or the amount of interactive performance allocated to a partition and the new
value is still within the minimum or maximum values for that partition, only the
affected partition needs to be IPLed. If the new value allocated requires the
minimum or maximum values to be changed, this requires an IPL of the
primary partition before it comes into effect.
If any partition has a changed "Connect to System Inter-Partition
OptiConnect" parameter, then an IPL of the primary partition is required
before this change becomes activated.
If a partition requires an IPL to put into effect some changes that were made
to the processing resources, a "<" symbol is shown next to that partition on
44
Slicing the AS/400 with Logical Partitioning: A How to Guide
the Work with Partition Configuration screen. If these changes are reversed
before an IPL, the symbol is removed from the display.
Work with Partition Configuration
System: SYSTEMA
Type option, press Enter.
1=Change partition name
2=Change partition processing resources
3=Add I/O resources
4=Remove I/O resources
5=Change bus ownership type 6=Select load source resource
Partition
Option Identifier Name
0
PRIMARY
1
LPAR2
F3=Exit
<
F11=Work with partition status
F12=Cancel
F23=More options
Figure 18. Work with Partition Configuration Display with Changes Pending
3.6 Saving and Restoring Partitions
Introducing partitions into a customer environment may add new steps into
the customers save and restore process. It should also initiate a review of the
entire save and restore strategy.
There is no way to perform a single save or restore for a physical system.
Each partition needs to be treated separately, but it is possible to perform
multiple saves or restores in different partitions at the same time, assuming
they each have their own save and restore device. In principle, the process of
saving or restoring logical partitions is the same as saving or restoring a
system without logical partitioning.
You can use the Save Storage (SAVSTG) command to save partitions data.
Remember that this command causes the system to power down and start
the system again as if PWRDWNSYS RESTART(*YES) was specified. If you
are going to use this command on the primary partition, make sure that all
secondary partitions are powered off before you start.
Operational Considerations
45
3.6.1 Saving the Logical Partition Configuration
You cannot save the configuration data for logical partitions to a removable
media device. The system automatically maintains the configuration data for
all logical partitions created on a physical system on the load source disk of
each partition.
When the system is IPLed, it validates the configuration available on the load
source disk of the primary partition. If this is not available, it searches the
other load source disks and prompts you to select the configuration that
should be used.
On a scratch installation of the primary partition, the partition configuration
must be recovered before the recovery of the disk configuration. For more
information on the steps that are required, refer to Backup and Recovery,
SC41-5304.
Not saving the partition configuration with the save commands allows data to
be restored to a system whether it has logical partitions. A full system save
can be used to restore to:
•
•
•
•
The same logical partition
A different logical partition on the same system
A different logical partition on a different system
A system that does not have any partitions at all
The save and restore process uses the machine serial number to determine if
an object is being restored to the system from which it was saved. Starting
with V4R4M0, a save or restore uses the serial number and partition number
to determine this. Objects that were saved prior to V4R4M0 and are restored
to a V4R4M0 system assumes that they are being restored to the same
system if the serial number matches.
3.6.2 Saving Data from Logical Partitions
All save operations operate on the logical partitions as if they were separate
systems. Using the save commands available, a customer may save their
logical partition as they do today. The save will contain only the information
from that logical partition.
3.6.3 Restoring Data to Logical Partitions
The restore commands available to customers may restore their logical
partition data the same way as they would restore data to their system today.
The commands that are used place the data into the partition from where the
restore command is issued.
46
Slicing the AS/400 with Logical Partitioning: A How to Guide
3.7 System Values Affected by Logical Partitioning
Every partition operates independently of the others. Most system values are
unique to that partition. However, there are some system values that can only
be changed from the primary partition. The value to which they are set
influences how the secondary partitions will behave in certain circumstances.
3.7.1 QPRCMLTTSK — Processor Multi-Tasking
The QPRCMLTTSK system value allows you to turn on and turn off the
processor multi-tasking capability. If enabled, more than one set of task data
will reside in each CPU. Some workloads may experience increased
performance due to caching implications. This system value setting affects all
processors in all partitions.
3.7.2 QPWRRSTIPL — Automatic IPL after Power Restored
The QPWRRSTIPL value specifies whether the system should automatically
IPL when utility power is restored after a power failure. It only affects the
primary partition. Whether a secondary partition is IPLed at the same time as
the primary partition depends on the secondary partition's configuration value
for System IPL Action.
If the secondary partition System IPL Action is set to IPL, the partition
automatically performs an IPL when the primary partition is IPLed. This is set
from the Work with Partition Status display on the primary partition.
Operational Considerations
47
Work with Partition Status
System:
Type options, press Enter.
13=IPL partition on system IPL
20=Display system type/model
22=Force Main Storage Dump
Partition
Opt Identifier Name
__
0
PRIMARY
__
1
LPAR2
__
2
LPAR3
IPL
Source
B
B
B
SYSTEMA
14=Hold partition on system IPL
21=Force Dedicated Service Tools
34=Force CPM or MSD IPL retry
IPL
Mode
Normal
Manual
Manual
F3=Exit F5=Refresh
F11=Work with partition configuration
State
On
On
Off
Sys IPL
Action
IPL
IPL
HOLD
Reference
Codes
11
11
11
F10=Monitor partition status
F12=Cancel F23=More options
Figure 19. Work with Partition Status Display
3.7.3 QRMTIPL — Remote Power On and IPL
The QRMTIPL value specifies whether remote power on and IPL can be
started over a telephone line. It only affects the primary partition. Whether a
secondary partition is IPLed at the same time as the primary partition
depends on the secondary partition's configuration value for System IPL
Action.
If the secondary partition System IPL Action is set to IPL, the partition
automatically performs an IPL when the primary partition is IPLed. See
Figure 19 for details on setting this parameter.
3.7.4 QUPSDLYTIM — Uninterruptible Power Supply Delay Time
The QUPSDLYTIM value specifies the amount of time that elapses before the
system automatically powers down following a power failure. When a change
in power activates the Uninterruptible Power Supply (UPS), messages are
sent to the UPS message queue (the system value QUPSMSGQ). This
system value is only meaningful if your system has a battery power unit or
has an uninterruptible power supply.
48
Slicing the AS/400 with Logical Partitioning: A How to Guide
Once the primary partition receives notification of a power failure, it notifies
the secondary partition that such a condition exists. The messages are sent
to the UPS message queue specified in the QUPSMSGQ on the secondary
partitions. The value specified in the primary partition is the value that the
system will use for all partitions.
It is important that you monitor for power failures across all partitions, and not
just primary partitions. This ensures that the secondary partitions (which may
be running a production workload) perform a normal shutdown in the event of
a power failure.
Note
If any of the previously described system values are displayed or
retrieved from a secondary partition, the value always reflects what was
specified in the primary partition.
3.7.5 Other Related System Values and Commands
The following list of system values and commands are not enforced by the
operating system. However, their values can impact operating partitions.
• QAUDENDACN — Auditing end action
Contains the action that should be taken by the system when audit records
cannot be sent to the auditing journal because of errors that occur when
the journal entry is sent. If this is set to *PWRDWNSYS, the system ends
should the attempt to send the audit data to the security audit journal fails.
It ends with a B900 3D10 system reference code.
If this system value is set to *PWRDWNSYS in the primary partition, then
this would effect all secondary partitions should a problem be
encountered.
• QCMNARB — Communication arbiters
Contains the number of communication arbiter system jobs that are
available to process work for controllers and devices. If set to *CALC (the
shipped value), the operating system calculates the number of
communication arbiter system jobs to be equal to the number of
processors on the system.
On a partitioned system, the number of jobs started are equal to the
number of processors on the physical system, not the number assigned to
that partition. We recommend that this system value be set manually.
Operational Considerations
49
• QIPLDATTIM — Date and time to automatically IPL
Specifies a date and time when an automatic IPL should occur. This
system value can be set independently in each partition. If the primary
partition is powered down at the time that an automatic IPL should occur in
a secondary partition, the IPL will not occur. When the primary partition
performs an IPL, the secondary partition is IPLed if its IPL date and time
are past due. The secondary partition performs an IPL even if it was
configured with a System IPL Action of Hold.
The power on and off schedule that can be set from the POWER menu
also uses the system value QIPLDATTIM. This function works the same
way as described in the previous paragraph.
• QMODEL — System model number
This is the number or letter used to identify the model of the system. You
cannot change QMODEL, but the four-character value can be displayed or
retrieved in user-written programs. The system model number is the same
in each partition on a system.
• QPRCFEAT — Processor feature
This is the processor feature code level of the system. You cannot change
QPRCFEAT, but the four-character value can be displayed or retrieved in
user-written programs. The processor feature system value is the same in
each partition on a system.
• QSRLNBR — System serial number
This value cannot be changed. It is retrieved from the data fields by the
system when installing the OS/400 licensed program. You can display
QSRLNBR, or you can retrieve this value in user-written programs. The
system serial number is the same in each partition on a system.
• CHGIPLA Command
The Change IPL Attributes (CHGIPLA) command allows you to change the
settings of attributes that are used during the initial program load (IPL).
The "Hardware diagnostics (HDWDIAG)" attribute on this command
specifies whether certain hardware diagnostics should be performed
during the IPL. The list of diagnostics is pre-determined by the system and
cannot be modified by the user.
Hardware diagnostics can only occur when the primary partition is IPLed.
This attribute cannot be changed from a secondary partition. Attempting to
change this attribute in a secondary partition will result in message
CPF18B4 being issued.
50
Slicing the AS/400 with Logical Partitioning: A How to Guide
3.8 Security Considerations
It is possible to have different OS/400 security levels (QSECURITY) in each
partition and a different security policy implemented by the customer for each
partition. One partition may have security down to the object level, but
another may have all objects secured by group profiles.
A user profile is required for each partition on a physical machine.
System-supplied user profiles will exist on each partition after the operating
system is installed. However, there is no dependency between partitions for
these profiles.
It is possible for user profiles, for example QSRV, to have different passwords
for each partition on the physical system. Typically, most maintenance
requiring access by an IBM service representative is done from the primary
partition. All QSRV passwords should be made available in case they are
required.
When configuring partitions, it is possible to specify that Interpartition
OptiConnect should be enabled. Any changes made to this parameter after
the initial creation requires a full system IPL. For details on how to restrict
user authority to OptiConnect, refer to the manual OptiConnect for OS/400,
SC41-5414.
3.9 Date and Time Processing between Partitions
Each partition has its own system date and time as set by the QDATE and
QTIME system values. There is only one clock on each physical system. This
stores the value of QDATE and QTIME from the primary partition. To ensure
that the secondary partitions date and time are stored correctly, each
secondary partition maintains a delta value to the system clock. They use this
to set the partition date and time when the partition is IPLed.
If the date or time is changed in the primary partition, the service processor
updates the delta in each secondary partition to maintain the correct date and
time for the secondary partitions. Each time that a change is made to the date
or time in a secondary partition, the virtual service processor updates the
delta accordingly.
If a hardware problem occurs in the primary partition that requires the date
and time to be entered on the next IPL, the system automatically signals all
secondary partitions that their date and time values are invalid. On the IPL of
Operational Considerations
51
the secondary partitions, the correct date and time for each partition must be
entered when prompted to do so.
It is possible to display the reference code history of the secondary partitions
from the primary partition. The date and time displayed indicate when the
primary partition detected the reference code. They are not the date and time
from the secondary partition settings.
Display Secondary Partition Reference Code History
System:
Secondary partition(s) to display . . . . . *ALL *ALL, 1-1
Number of reference codes to display . . . 200 1-200
Partition
Identifier Name
1
LPAR2
Reference Codes
11
11 C900 2F00
11 C900 2C25
11 C900 2C20
11 C900 2C40
11 C900 2C10
11 C900 2B40
11 C900 2B30
11 C900 2B10
11 C900 2AC0
11 C900 2AB0
11 C900 2AA5
11 C900 2AA4
Date
01/24/99
01/24/99
01/24/99
01/24/99
01/24/99
01/24/99
01/24/99
01/24/99
01/24/99
01/24/99
01/24/99
01/24/99
01/24/99
SYSTEMA
Time
20:56:45
20:56:45
20:56:44
20:56:44
20:56:44
20:56:13
20:56:12
20:56:11
20:56:02
20:56:01
20:56:01
20:56:00
20:56:00
More...
F3=Exit F5=Refresh
F9=Include reference code detail
F6=Print
F12=Cancel
Figure 20. Display Secondary Partition Reference Code History
3.10 Main Storage Dump Processing
The main storage for each partition is specified on the Create Partition or
Change Partition Processing Resources display. The primary partition
requires a minimum of 256 MB of main storage. Secondary partitions require
a minimum of 64 MB. Creating a partition with only the minimum amount of
main storage specified allows only a very limited workload to run in that
partition.
Main storage can be allocated in 1 MB increments over and beyond this
minimum. This can result in one physical main storage card being allocated to
two partitions. This is not a problem.
52
Slicing the AS/400 with Logical Partitioning: A How to Guide
For completeness purposes, the main storage dump (MSD) process is
explained in the following steps to show how the system copes with this type
of configuration:
1. Once a fail condition is detected, the secondary logical partition sends a
signal to the virtual service processor.
2. In the primary partition, the signal to the virtual service processor also
results in the SRC being written to the log.
3. All IOPs that are dedicated to the failing partition are reset.
4. The IOP that has the load source DASD attached to it is IPLed by the
primary partition.
5. The secondary partition copies 128 MB of memory to the load source
DASD and continues to IPL to DST with that 128 MB of memory.
6. The secondary partition performs MSD processing and displays the MSD
screens.
7. Once MSD processing is complete, the partition is IPLed.
A failure in the primary partition will cause all secondary partitions to end
abnormally. Extra options were added to the main storage dump process
display to allow a specific main storage dump, or all of them, to be selected.
3.11 Managing the Hardware Resource Configuration
You can display the hardware resource configuration from the DST, SST, and
logical partition displays. Where you view the configuration from determines
the level of detail that is displayed. This section explains the information
available from each display.
3.11.1 Hardware Resource Information from DST and SST
Use the Hardware Service Manager within DST or SST to display the
installed hardware resources. The hardware service manager is typically
used by service representatives when performing hardware maintenance on
the system.
The processors, main storage, and bus adapters that are displayed are for
the entire system, not just the partition from where the option is being viewed.
The IOP resources displayed are those currently reporting to the partition
from where the option is being viewed.
Failed and non-reporting resources can also be displayed. These resources,
at one time, were reported as operational for that partition.
Operational Considerations
53
It is possible to print the configuration displayed by selecting F6 (Print
configuration) from the Hardware Service Manager initial menu (see Figure
21).
Hardware Service Manager
Attention: This utility is provided for service representative use only.
System unit . . . . . . . :
Release . . . . . . . . . :
9406-720 10-394TM
V4R4M0 (1)
Select one of the following:
1.
2.
3.
4.
5.
6.
7.
8.
Packaging hardware resources (systems, frames, cards,...)
Logical hardware resources (buses, IOPs, controllers,...)
Locate resource by resource name
Failed and non-reporting hardware resources
System power control network (SPCN)
Work with service action log
Display label location work sheet
Device Concurrent Maintenance
Selection 2
F3=Exit
F6=Print configuration
F9=Display card gap information
F10=Display resources requiring attention
F12=Cancel
Figure 21. Hardware Service Manager — Initial Menu
54
Slicing the AS/400 with Logical Partitioning: A How to Guide
Logical Hardware Resources on System Bus
System bus(es) to work with . . . . . .
Subset by . . . . . . . . . . . . . . .
*ALL *ALL, 1- 2
*ALL *ALL, *STG, *WS, *CMN, *CRP
Type options, press Enter.
2=Change detail
4=Remove
5=Display detail
6=I/O Debug
8=Associated packaging resource(s)
9=Resources associated with IOP
Opt Description
System Bus
Multiple Function IOP
System Bus
Virtual System Bus
Virtual Bus Adapter
Type-Model
* < 9164-001
268A-000
Status
Operational
Operational
Operational
Operational
Operational
Resource
Name
SPD01
CMB01
SPD02
SPD03
SOC01
F3=Exit
F5=Refresh
F6=Print
F8=Include non-reporting resources
F9=Failed resources
F10=Non-reporting resources
F11=Display serial/part numbers
F12=Cancel
Figure 22. Hardware Service Manager — Sample Resources Display
3.11.2 Hardware Resource Information from Logical Partitions
You can only access I/O resource information from within the logical
partitioning displays. The amount of information that is available depends on
whether you are viewing it from the primary or secondary partition. The I/O
resource information for each partition is rebuilt when a partition IPLs.
If an IOP within a tower is replaced using the concurrent maintenance facility,
the I/O resources are not updated properly until the partition is IPLed. If a
partition is currently powered off, the I/O resources displayed reflect the
configuration at the time the partition was powered off.
When switching I/O resources between partitions, a time delay may exist
before the resource is displayed as allocated to the partition. This is a result
of the time it takes the IOP to IPL and update the resource information.
3.11.2.1 Information Available from the Primary Partition
To display the allocated I/O resources for all partitions in partition ID
sequence, perform the following steps from the Work with System Partitions
initial menu:
1. Select option 1 (Display partition information).
2. Select option 3 (Display allocated I/O resources).
Operational Considerations
55
Display Allocated I/O Resources
System: RCHASM25
System partition(s) to display . . . . *ALL *ALL, 0-1
Level of detail to display . . . . . . *ALL *ALL, *BUS, *IOP, *IOA, *DEV
Par
ID Description
0 PRIMARY Partition
System Bus
Multiple Function IOP
Communications IOA
Communications Port
Communications IOA
Communications Port
Communications Port
Virtual Port
Communications IOA
Workstation IOA
Display Station
*<
Serial
Type-Model Number
9406-720
10-394TM
00-0000000
9164-001
10-8285313
2720-001
53-8302889
2720-001
53-8302889
2850-011
10-8282050
2724-001
10-8215054
2838-001
10-8316090
6B00-001
10-8282050
285A-003
10-8301085
266C-001
53-8302889
3487-0HC
00-*******
Part
Number
0000087G0725
0000090H9189
0000090H9189
0000061H0127
0000044H7571
0000021H5458
000008193654
0000090H9212
0000090H9189
More...
* Indicates load source.
F3=Exit F5=Refresh F6=Print
F11=Display partition status
F10=Display logical address
F12=Cancel
Figure 23. Display Allocated I/O Resources from Primary Partition
To display the available I/O resources that can be allocated to the primary
partition, perform the following steps from the Work with System Partitions
initial menu:
1. Select option 1 (Display partition information).
2. Select option 4 (Display available I/O resources).
The available resources are attached to the system and are not allocated to
another partition. For example, you may have a tape drive that will be shared
across all of the partitions.
To display the total system I/O resources, perform the following steps from
the Work with System Partitions initial menu:
1. Select option 1 (Display partition information).
2. Select option 5 (Display system I/O resources).
56
Slicing the AS/400 with Logical Partitioning: A How to Guide
Display System I/O Resources
System: RCHASM25
Level of detail to display . . . . . . *ALL *ALL, *BUS, *IOP, *IOA, *DEV
I/O Resource
Description
System Bus
Multiple Function IOP *<
Communications IOA
Communications Port
Communications IOA
Communications Port
Communications Port
Virtual Port
Communications IOA
Workstation IOA
Display Station
Display Station
Multiple Function IOA
Disk Unit
Serial
Type-Model Number
00-0000000
9164-001
10-8285313
2720-001
53-8302889
2720-001
53-8302889
2850-011
10-8282050
2724-001
10-8215054
2838-001
10-8316090
6B00-001
10-8282050
285A-003
10-8301085
266C-001
53-8302889
3487-0HC
00-*******
3487-0HC
00-*******
2741-001
10-8308021
6713-074
68-0171DE9
Part
Number
0000087G0725
0000090H9189
0000090H9189
0000061H0127
0000044H7571
0000021H5458
000008193654
0000090H9212
0000090H9189
0000091H0271
59H6611
More...
F3=Exit F5=Refresh F6=Print
F11=Display partition status
F10=Display logical address
F12=Cancel
Figure 24. Display System I/O Resources Display
It is possible to print the system I/O resource information by selecting the F6
key from the display.
3.11.2.2 Information Available from the Secondary Partition
If option 3 is selected from the Display Partition Information menu on the
secondary partition, only the allocated I/O resources for the secondary
partition are displayed.
If option 4 is selected from the Display Partition Information menu on the
secondary partition, only the available I/O resources that can be allocated to
the secondary partition are displayed.
3.11.3 Removing Hardware Resources from Logical Partitions
Whether you are creating logical partitions for the first time or switching some
hardware resources between partitions, it is important to understand how the
resources should be removed from the current partition. You want to ensure
that the hardware resource table is updated correctly.
Operational Considerations
57
Remove I/O Resources
System: RCHASM25
Level of detail to display . . . . . . *ALL *ALL, *BUS, *IOP, *IOA, *DEV
Partition identifier . . . . . . . . . . . . . : 0
Partition name . . . . . . . . . . . . . . . . : PRIMARY
Type options, press Enter.
1=Remove 2=Remove and clear hardware resource(s)
I/O Resource
Opt Description
System Bus
Multiple Function IOP
Communications IOA
Communications Port
Communications IOA
Communications Port
Communications Port
Virtual Port
*<
Serial
Type-Model Number
00-0000000
9164-001
10-8285313
2720-001
53-8302889
2720-001
53-8302889
2850-011
10-8282050
2724-001
10-8215054
2838-001
10-8316090
6B00-001
10-8282050
* Indicates load source.
F3=Exit F10=Display logical address
Part
Number
0000087G0725
0000090H9189
0000090H9189
0000061H0127
0000044H7571
0000021H5458
000008193654
More...
F12=Cancel
Figure 25. Remove I/O Resources Display
If the allocation of the hardware resource is going to be permanently removed
from that partition, remove it by using option 2. This ensures that the
hardware resource tables are updated correctly.
If the hardware resource is going to be used in this partition again, then
remove it by using option 1. This ensures that the hardware resource names
previously assigned by the system are kept and reused later when the
hardware resource is reallocated to the partition. While the hardware
resource is not allocated to the partition, it will have a status of "Non
Reporting" when viewed from the Hardware Service Manager.
It is possible to remove non-reporting hardware resources by using the
Hardware Service Manager in DST and SST. Or, option 4 can be used from
the logical partitioning Recover Configuration Data menu. This should only be
done by experienced service personnel. Incorrect use of these options may
result in valid resource names being incorrectly removed.
3.11.4 Printing the System Configuration
We recommend that, once logical partitions are initially setup and every time
the configuration changes, you print the system configuration and keep it with
the service and disaster recovery documentation.
58
Slicing the AS/400 with Logical Partitioning: A How to Guide
Follow these steps to print the system configuration:
1. From the primary partition, start SST or DST.
2. From SST, select option 5 (Work with system partitions). From DST, select
option 11 (Work with system partitions), and press Enter.
3. Select option 1 (Display partition information).
4. Select option 5 (Display system I/O resources).
5. Press F6 to print the configuration.
6. Select either option 1 (132 characters wide) or 2 (80 characters wide).
Press Enter to print to a spooled file.
7. Press F12 to return to the Display Partition Information screen.
8. Select option 2 (Display partition processing configuration).
9. Press the Print Screen key to print the configuration to a spooled file.
10.Press F10 (Display main storage information).
11.Press the Print Screen key to print the configuration to a spooled file.
12.Press F10 (Display interactive information).
13.Press the Print Screen key to print the configuration to a spooled file.
14.Press F10 (Display processor identifiers).
15.Press the Print Screen key to print the configuration to a spooled file.
16.Print the spooled files that were generated.
3.12 IPL Process
When the primary partition is IPLed, it will not start the IPL of the secondary
partitions, unless specific criteria is entered. First, the serial number must be
validated. Second, the system password checked to ensure it has not
expired.
If the secondary partition was configured with a System IPL Action of IPL, the
primary partition signals the secondary partition to start the IPL. The IPLs of
all secondary partitions are started at the same time, and run concurrently.
During this IPL process, all hardware checks are bypassed. Some of these
were already completed by the IPL process of the primary partition. A normal
IPL process of the secondary partition is significantly shorter than the primary
partition.
3.13 Licensed Key Management
A licensed authorization code must be installed on your system. If the system
prompts you to enter this key, then it will allow you to bypass the prompt and
IPL the system for a maximum of 70 days. After the 70-day grace period,
OS/400 will not IPL until a valid key is entered.
Operational Considerations
59
On a partitioned system, the licensed authorization code supplied by IBM has
to be entered into each individual partition since each partition has its own
copy of OS/400. The same license key can be entered for all logical partitions
that are defined on the system.
There are two licensed keys. One is for the system password. The other is for
OS/400 licensed users. Licensed keys for all other licensed programs and
applications also have to be entered on all partitions. There is no automatic
transfer of license information between the partitions.
3.14 Work Management
Take the following information into account when adjusting the work
management setup for each partition.
3.14.1 Machine Pool
Every AS/400 system automatically reserves an amount of the machine pool
for internal use. The system enforced minimum value varies depending on
the main storage size of the machine. The total current reserved size can be
seen when using the Work with System Status (WRKSYSSTS) command.
When creating or changing a partition, the minimum and maximum values
used for main storage can have a direct influence on the machine pool
reserve size.
When the system is IPLed, it calculates the amount of main storage that is
required to support the internal hardware page table for that maximum value.
Then, the system adds this amount to the reserved figure. If the maximum
value is set to an unrealistic value for that partition, the main storage is
unnecessarily reserved in the machine pool and unavailable for use by other
tasks and partitions.
For example, a system has 20 GB of main storage with three partitions. If
each partition has the maximum amount of main storage set to 20 GB, the
system reserves 512 MB in the machine pool for the hardware page table. It
is obvious that it would not be possible for each partition to have 20 GB of
main storage allocated at the same time. In this case, some of the 512 MB
reserved in the machine pool is wasted.
At V4R4M0, the minimum and maximum values primarily affect when an IPL
is required. If you change the amount of main storage allocated to a partition
and it is within the minimum and maximum values, then only an IPL of that
partition is required to make the change active. If the change to the amount of
main storage allocated also requires a change to either of the minimum and
60
Slicing the AS/400 with Logical Partitioning: A How to Guide
maximum values, then this requires an IPL of the primary partition to activate
this change.
Note
The main storage totals seen on the WRKSYSSTS display may not equal
the total for the logical partition. The difference is related to the size of the
hardware page table. To reduce the difference, the maximum value for the
partition should be lowered to be closer to the actual amount allocated.
3.14.2 QPFRADJ
The systems performance adjuster examines configuration related objects as
one of the many inputs. It also makes performance adjustments accordingly.
We recommend that all obsolete configuration related objects be removed.
This is especially necessary if the partitioned system was originally a single
N-way machine, and the creation of partitions made a number of
configuration objects obsolete in the primary partition.
Operational Considerations
61
62
Slicing the AS/400 with Logical Partitioning: A How to Guide
Chapter 4. Change Management for a Logical Partition
A partitioned system may need to be changed. As long as your system
supports the criteria for partitioning, you can change the setup of partitions
inside these criteria. For more information about the criteria, see Chapter 2,
“Planning Considerations” on page 17. For a detailed description on how to
set up or do changes to your partitioned system, you can also refer to the
articles found on the LPAR home page at: http://www.as400.ibm.com/lpar
• Planning for and Setting Up Logical Partitions
• Managing Logical Partitions
• Backing up and Recovering Logical Partitions
It is important to plan what you will do before you start. Taking the time to plan
will save steps and avoid trouble during the partition setup. It is also
important to understand that partitioning a system or changing the partition
setup, using dedicated service tool (DST) or system service tool (SST), may
seriously impact your system. Therefore, only specially dedicated persons
should be allowed to do this. See "Controlling your Partitions with DST and
SST" in the Managing Logical Partitions article for specific details on what
DST and SST can do for logical partitions.
Be careful to read the LPAR planning articles before you make any changes
to a partition configuration. Otherwise, this may cause you to perform an
Initial Program Load on your system. It may also cause problems for normal
work going on the system. Changing your partition setup is done from DST in
your primary partition.
4.1 Managing Partitions
Changing logical partitions consists of several tasks. You may change:
• Logical partition processing resources
• Logical partition configuration
These steps are documented in detail in the Planning for and Setting Up
Logical Partitions article.
4.2 Managing User Applications in a Logical Partition
Each partition is seen as a separate system in itself. There are no interfaces
between the different partitions, other than what you have between separate
applications in separate machines. That means normal communication setup.
© Copyright IBM Corp. 1999
63
The high-speed internal OptiConnect that can be defined in logically
partitioned systems will, from an application point of view, be seen as if it is
between separate boxes. See Chapter 7, “Inter-Partition Communication” on
page 133.
Installing new user applications or maintaining an old one should not involve
any special considerations working in a partition compared to a single
system. Any application that you use now to maintain or distribute changes in
applications to other systems, will work for maintaining or distributing
changes in applications in single-system partitions.
4.2.1 Distributing Changes through Management Central
If you have changes that you want to distribute to other systems or partitions,
you may use Management Central to do this. Management Central provides
support for managing multiple AS/400 systems and is part of AS/400
Operations Navigator. AS/400 Operations Navigator allows you to manage
your AS/400 server through a graphical interface on your Personal Computer
(PC). AS/400 Operations Navigator is part of AS/400 Client Access Express,
which is free with V4R4 of OS/400. When you install Client Access Express,
Management Central will not be automatically installed. You have to select to
install it. The interprocess communication between AS/400 systems (in this
case, partitions) is done with TCP/IP.
On your source system, which is your selected partition, you can create
packages with files and objects that can be distributed to target systems, or
other partitions. You may also specify actions, or a command, to be executed
on your target system when this package is successfully sent. This package
may be scheduled to be sent later.
This means that you can:
• Distribute a batch input stream and run it on a remote system
• Distribute a set of programs and start your application
• Distribute a set of data files and run a program that uses that data
If you have partitions running the same applications, you may set up your
applications to periodically send changes or updates to these partitions. The
packages may consist of files and objects or they may be packed into
savefiles. Management Central then takes care of distributing and handling
these packages.
64
Slicing the AS/400 with Logical Partitioning: A How to Guide
Figure 26. Sending a Package in Management Central
System Manager for AS/400 (5769-SM1) may also be used for distributing
fixes between partitions. Independent software providers may use this
product to distribute application fixes in the same way as normal OS/400
fixes. System Manager for AS/400 uses SNA for communication.
4.3 Program Temporary Fix (PTF) Management
The primary partition normally has the Electronic Customer Support line
installed on the multifunction input/output processor (MFIOP). This is used for
ordering PTFs for the primary partition. If communication IOPs are installed in
the different partitions, you may use communication lines to define Electronic
Customer Support lines for each partition.
If there are no communication lines, you have to copy the PTFs to media from
the primary partition. Then, you must install it on the secondary partition from
the saved media.
When installing PTFs, follow the steps that are described in "Installing
temporary fixes on a partitioned system" in the Managing Logical Partition
article.
Change Management for a Logical Partition
65
4.3.1 Using Management Central in PTF Distribution
If you defined TCP/IP in the different partitions, you may use Management
Central to deal with the fixes. Your primary system may be specified as your
central system for Management Central. By using this as your endpoint
system, you can copy PTFs from such media as tape or CD-ROM to your
system.
In Configuration Services, you work with fixes on your system. Here you may
install fixes for all or specific products. You may install them at once or
schedule when you want to install them.
To distribute the PTFs to a partitioned system, keep the fixes in savefiles on
your primary partition. You can use Management Central to copy the fixes
from media, or order the fixes through the Electronic Customer Support line.
Using Management Central, perform the following steps:
1. Right click on your system under Endpoint Systems.
2. Click on Fixes—>Compare and Update (Figure 27 on page 67).
Now, you are in the wizard for setting up the transfer of fixes to other
systems, which, here, is partitions.
3.
4.
5.
6.
Select the model system that is your primary system.
Select available systems or groups that are your partitions.
Select each of the products for which you are choosing to work with fixes.
Perform one of the following tasks:
• Compare the results. Compare fixes on the primary partition against
the secondary partition to see which fixes are missing on each
secondary partition.
• Send missing fixes from the primary partition to the secondary partition
that you have chosen.
• Send and install missing fixes on the secondary systems.
7. Run what you have chosen now or schedule it for later. If the fixes need an
Initial Program Load, you may specify it now.
Putting on new cumulative packages with fixes is not handled through
Management Central. Complete the instructions following the cumulative
packages. You should install the cumulative packages using the Program
Temporary Fix menu. This means you should use the CD-ROM installation
media to install these package fixes. Since Management Central uses
savefiles to distribute fixes, you may copy single fixes from the cumulative fix
media. Then, install the fixes using Management Central.
66
Slicing the AS/400 with Logical Partitioning: A How to Guide
Figure 27. Sending Fixes Using Management Central
4.3.1.1 Why Manage Fixes with Management Central
One of the key benefits of Management Central is that it makes managing
multiple systems, in this case partitions, as easy as managing a single
system. Managing fixes is simplified with the tools that are included with
Management Central. These tools include several wizards to guide you
through these tasks:
•
•
•
•
•
Installing fixes
Permanently installing fixes
Sending and installing fixes
Uninstalling fixes
Comparing and updating fixes
For example, to install multiple fixes, you select the fixes from a list and start
the install wizard. You may find the Compare and Update wizard to be very
beneficial. The wizard compares the fix level of a single system or multiple
systems to a model system and makes them the same. Using the collected
inventory, you can have the system compare the fix levels for a group of
systems to a model system, send the save files of the selected fixes from the
source system, and install the fixes.
Change Management for a Logical Partition
67
4.4 Managing Security in Logical Partition
The security-related tasks you perform on a partitioned system are the same
as on a system without logical partitions. However, when you create logical
partitions, you work with more than one independent system. Therefore, you
have to perform the same tasks on each logical partition instead of just once
on a system without logical partitions. Here are some basic rules to
remember when dealing with security on logical partitions:
• Add users to the system one logical partition at a time. You need to add
your users to each logical partition that you want them to access.
• Limit the number of people who have authority to go to Dedicated Service
Tools and System Service Tools on the primary partition. Refer to the table
in "Controlling your partitions with DST and SST" in the Managing Logical
Partitions article for more information on DST and SST.
• Secondary partitions cannot see or use main storage and disk units of
another logical partition.
• Secondary partitions can only see their own hardware resources.
• The primary partition can see all system hardware resources in the Work
with System Partitions displays of Dedicated Service Tools and System
Service Tools.
• The operating system still only sees resources available to its own
partition.
To manage users and user groups in your partitions, you can use Operations
Navigator in AS/400 Client Access Express. From Users and Groups option,
you can see all users and groups on an AS/400 system. From there you can
manage AS/400 users and groups. For each AS/400 system to which you are
connected, you can work with all users, groups, and users not in a group. You
can perform the following actions with Users and Groups:
•
•
•
•
•
•
Create and change users
Add users to groups
Copy users to other AS/400 systems (partitions)
Delete users
Create or change a group profile
Drag and drop users and groups
For example, you can drag a user to Users Not in a Group to remove the
user from all groups. You can also drag a user to a group name to add the
user to the group, add a user to another system (partition), or move the
user to the group.
68
Slicing the AS/400 with Logical Partitioning: A How to Guide
Chapter 5. Problem Management for a Logical Partition
If you encounter problems on a partitioned AS/400 system, you should first
try to determine if they are general AS/400 problems, or problems specific to
logical partitions. Refer to your AS/400 manuals on problem determination to
evaluate whether your problem is a general AS/400 problem.
For recovery information in regard to logical partitions, refer to the Backing
Up and Recovering Logical Partition article. This article is available from the
LPAR home page at: http://www.as400.ibm.com/lpar
For other partition specific problems, you can try any of these options:
• Work with logical partition error messages and reports that appear on your
console.
• Work with main storage dumps.
• Use remote service to get service on your system through a modem.
• Find out when you should power off and on a domain.
• Find out when you need to reset a disk unit input/output processor.
These options are also documented in "Troubleshooting logical partitions" in
the Managing Logical Partitions article (also available from the LPAR home
page).
5.1 Problem Determination for a Logical Partition
If the system experiences logical partition errors, it notifies you in one of two
ways:
• Through an error message that appears at the bottom of the display
• Through the Logical Partitioning Error Report display that appears on the
system console
From the primary system console, you may look at System Reference Codes
and Initial Program Load codes for all partitions. From the secondary partition
console, you may look at that partition’s System Reference Codes and Initial
Program Load codes for secondary only. You may choose to only see the
main codes or use a command key to display all codes. See the "Displaying
reference code history for secondary partitions" in the Managing Logical
Partitions article.
© Copyright IBM Corp. 1999
69
5.2 Reporting to IBM
Each partition is like a single system. You may report problems from the
problem log. This means that there has to be an Electronic Customer Support
line defined in the partition. Through communication, you may set up one
partition to work as a "hub" for the other partitions.
The Service Director is a tool that runs on the AS/400 system to supervise
any problems encountered. It may send for Program Temporary Fixes or send
messages to IBM informing of problems in the system. The main purpose is
to encounter and handle problems before they get into serious problems for
the system. The Service Director runs in each partition as if it is a single
system and uses an Electronic Customer Support line for the partition in
which it runs. Through communication lines between partitions, the Service
Director may be set up to use the Electronic Customer Support line in one
partition to communicate to IBM.
5.3 System Reference Codes (SRC)
The control panel on the system unit will show System Reference Codes
(SRC) for the system primary partition only. If no partitions are defined, the
entire system is viewed as a partition.
Each partition has its own Product Activity Log (PAL) and Service Action Log
(SAL) accessed through dedicated service tools (DST) or system service
tools (SST) that contain SRCs. The problem log also contains problem
determination information that you can access by using the Work with
Problems (WRKPRB) command.
Up to 200 of the most recent System Reference Codes for secondary
partitions can be found by using the primary partition’s console and accessing
either DST or SST. In addition, you must perform the following steps:
1.
2.
3.
4.
Enter the STRSST command on the command line.
Select option 5 (Work with system partitions).
Select option 1 (Display partition information).
Select option 7 (Display secondary partition reference code history).
Or use the following steps to gain access through DST:
1.
2.
3.
4.
70
Force dedicated service tools on the primary partition.
Sign on to DST.
Select option 11 (Work with system partitions).
Select option 1 (Display partition information).
Slicing the AS/400 with Logical Partitioning: A How to Guide
5. Select option 7 (Display secondary partition reference code history).
See Figure 28 for an example of the System Reference Code listing.
Secondary partition(s) to display . . . . . *ALL *ALL, 1-1
Number of reference codes to display . . . 200 1-200
This is screen.
Partition
Identifier Name
1
LPAR2
Reference Codes
11
11 C900 2F00
11 C900 2C25
11 C900 2C20
11 C900 2C40
11 C900 2C10
11 C900 2B40
11 C900 2B30
11 C900 2B10
11 C900 2AC0
11 C900 2AB0
11 C900 2AA5
11 C900 2AA4
F3=Exit F5=Refresh
F9=Include reference code detail
Date
01/26/99
01/26/99
01/26/99
01/26/99
01/26/99
01/26/99
01/26/99
01/26/99
01/26/99
01/26/99
01/26/99
01/26/99
01/26/99
Time
11:03:33
11:03:33
11:03:33
11:03:32
11:03:32
11:03:00
11:02:59
11:02:58
11:02:48
11:02:48
11:02:48
11:02:46
11:02:46
F6=Print
F12=Cancel
Figure 28. Display Secondary Partition Reference Code History
The reference code history log lists the SRCs. If the F9 (Include reference
code detail) key is used and extended, the reference codes 12 through 19 are
shown. See Figure 29.
Partition
Identifier Name
1
LPAR2
Reference Codes
11
11 C900 2F00
12 0DB0 0060
13 0000 0000
14 0000 0000
15 0000 0000
16 0000 0000
17 0000 0000
18 0000 0000
19 0000 0000
11 C900 2C25
12 0DB0 0060
13 0000 0000
F3=Exit F5=Refresh
F9=Exclude reference code detail
Date
Time
01/26/99 11:03:33
01/26/99 11:03:33
01/26/99 11:03:33
F6=Print
F12=Cancel
Figure 29. Include Reference Code Detail
Problem Management for a Logical Partition
71
5.3.1 Licensed Internal Code Log
The Licensed Internal Code log is accessed under Dedicated Service Tools or
System Service Tools. It has listings of major and minor codes that can be
used to compare with the Product Activity Log entries. Many times, PTFs are
written to fix problems that generate System Reference Codes or message
IDs that occur with specific Licensed Internal Code log entries. The latest
PTF for your release called the Preventive Service Planning (PSP), can be
downloaded electronically to your system using the Electronic Customer
Support line. Use the Send PTF Order command as shown here:
SNDPTFORD SF98VRM
Here, VRM is the version, release, and modification of OS/400 that is on your
system. If a PTF is found within the Preventive Service Planning to fix the
problem you encountered, follow the normal PTF procedures to procure, load,
and apply the fix.
5.4 Power Interruptions in the Logical Partition
Power interruptions are detected by the primary partition. The primary
partition then notifies the secondary partitions. This is an internal process that
needs no setup from the user. Each partition receives the message about
power interruption in the message queue specified in QUPSMSGQ. The
partition has to take actions as if it is a single system.
Note: The system value QUPSDLYTIM is a common system value for all
partitions.
5.5 Electronic Customer Support Lines
Only one Electronic Customer Support line is recommended per system.
However, if a partition will be used independently, it may be desirable for each
partition to have a separate service connection to order and receive PTFs.
There are three choices:
• A separate Electronic Customer Support communication line for each
partition.
• One Electronic Customer Support communication line for the primary
partition, which will act as a "hub" for the secondary partitions. Use
72
Slicing the AS/400 with Logical Partitioning: A How to Guide
Management Central in Operations Navigator in AS/400 Client Access
Express to distribute the fixes. See Chapter 4, “Change Management for a
Logical Partition” on page 63, about PTFs.
• One Electronic Customer Support communication line set up to be
switchable between secondary partitions.
In some places, it may be possible to receive PTFs over the Internet.
5.6 Logical Partition Configurations
Configuration data for all logical partitions is maintained on the load source
disk of each logical partition. You cannot save the configuration data to tape.
If a load source disk is initialized, the configuration data is copied from
another load source disk during the Initial Program Load. See the Backing up
and Recovering Logical Partitions article for recovery and working with
configuration data for logical partitions.
For partition configuration information, see "Displaying and printing logical
partition information" in the Managing Logical Partitions article.
5.6.1 Working with Configuration Data for Logical Partitions
The configuration data maintained on the primary partition is considered the
master copy. Problems can occur when configuration information on the
primary partition and a secondary partition conflict, or after you initialize disk
units.
The Recover Configuration Data menu (see Figure 30 on page 74) in DST
allows you to perform a different type of recovery on logical configuration
data.
Problem Management for a Logical Partition
73
Recover Configuration Data
System: SYSTEMA
Select one of the following:
1.
2.
3.
4.
5.
6.
7.
Recover primary partition configuration data
Update configuration data
Clear non-configured disk unit configuration data
Clear non-reporting logical partitioning resources
Accept load source disk unit
Copy configuration data to other side
Clear configuration data
Selection
F3=Exit
F12=Cancel
Figure 30. Recover Configuration Data
5.6.1.1 Recovering Primary Partition Configuration Data
Recovering primary partition configuration data applies only to primary
partitions and is part of a full system recovery. For more information, see the
Backup and Recovery manual, SC41-5304. After re-installing the Licensed
Internal Code on the primary partition, stop at the Initial Program Load. Or,
install the system display and go to Dedicated Service Tools. In the Work with
System Partitions menu, you will find Recover Configuration Data.
Select option 1. Doing so locates a nonconfigured disk unit that contains the
most recent configuration data for your system. The message No units found
with more current configuration data may display instead of a list of
resources. In this case, no unassigned disk units contain any appropriate
configuration information.
You should be able to select a resource with a date and time that fall within a
valid time frame for the configuration. When you confirm it, the system copies
the new information data to the primary partition load source and
automatically restarts the primary partition. The next time you perform an
Initial Program Load on your secondary partitions, the system updates their
logical partition configuration data from the primary partition.
74
Slicing the AS/400 with Logical Partitioning: A How to Guide
5.6.1.2 Updating Configuration Data
You should update configuration data only when instructed to do so by a
service representative. You can select the option to update the configuration
data for every logical partition by copying it from the primary partition to all
active logical partitions.
Important
Using this function when the configuration data on the primary partition is
incorrect can ruin your existing configuration.
5.6.1.3 Clearing Nonconfigured Disk Unit Configuration Data
When you move disk units between logical partitions or systems, you may
need to erase any old configuration data before the system can use the
logical partition again.
Clear the configuration data when disk units have these characteristics:
• A System Reference Code of B600 5311 is reported in the product activity
log against a nonconfigured disk unit.
• It is no longer a load source in its own logical partition.
• It is originally from a different logical partition where it was a load source,
or it is from a different system which had logical partitions.
Selecting this option takes effect immediately.
5.6.1.4 Clearing Nonreporting Logical Partitioning Resources
This procedure should not be run if any hardware is marked as "failed". It
should only be run when all system hardware is completely operational.
After you add, remove, or move hardware within a system with logical
partitions, you may have resources that are no longer available or that are
listed twice. Old entries may still exist in the configuration data. You can clean
up these listings so that all nonconfigured resources are erased from the
logical partition. You do this from Dedicated Service Tools display.
5.6.1.5 Accepting a Load Source Disk Unit
When configuration data on the load source of a logical partition is different
from what the system accepts, an error is logged in the Product Activity Log
with the System Reference Code B600 5311.
Problem Management for a Logical Partition
75
If you recently moved or added disk units either within the system or from
another system, they may still contain configuration data. If you do not want
to use this disk unit as the load source, you need to clear its configuration
data before proceeding. Otherwise, you may want to use this disk unit as your
new load source on the logical partition. In this case, you should select this
option from a DST display.
5.6.1.6 Copying Configuration Data to the Other Side
Your system may experience a disk read error of the logical partition
configuration data if the following reference codes appear when you restart
from one source but not another:
• 11 B193 4511
• 12 xxxx xx5D (where x equals any value 0 through 9 or A through F)
• 13 690A 2060
You can find these reference codes on the operations panel. Copy the data
from the functioning source to the faulty source with this procedure.
Note: Only attempt to do this when you are certain that the logical partition
restarts normally when using the other Initial Program Load source.
Ensure that the configuration information contained on this Initial Program
Load source is correct. Verify this by viewing your logical partition’s status. A
correct configuration would show your most recent configuration of logical
partitions. If it is correct, start copying the configuration data. If the
configuration is not correct, do not copy the data.
This option writes the logical partitioning configuration data stored on this
side of the load source disk unit to the other side. For example, it can write
from source A Licensed Internal Code to source B Licensed Internal Code, or
from source B Licensed Internal Code to source A Licensed Internal Code.
5.6.1.7 Clearing Configuration Data
If you need to clear all configuration data to return the system to a
nonpartitioned system again, use the clear configuration data option. In doing
so, the system will no longer be partitioned.
This option returns all hardware resources to the primary partition. However,
all user data within secondary partitions will be lost. Be careful and perform
adequate backups before selecting this option.
Restart the system in manual mode for the change to take effect. On the
Initial Program Load, go to Dedicated Service Tools and select the option to
clear configuration data from nonconfigured disk units on the Recovery
76
Slicing the AS/400 with Logical Partitioning: A How to Guide
Configuration Data display. Do this once for every former secondary partition
load source disk unit.
5.7 Recovery for Logical Partitions
Recovery for logical partitions falls within two main categories:
• Recovery of logical partition configuration data (this must be done one
logical partition at a time)
• Recovery of user data to a logical partition
This section explains these categories.
5.7.1 Recovery Considerations with Logical Partitions
As you go through the recovery steps outlined in the Backup and Recovery
manual, SC41-5304, keep these points in mind:
• Recovery occurs on a partition-by-partition basis. There is not a
system-wide recovery option for all logical partitions.
• Since each logical partition should have its own backup media, ensure
that you use the appropriate media for the logical partition you recover.
• Make sure that you recover the logical partition configuration data after
initializing the disk units of the primary partition.
• On a full system recovery, you must perform the following steps:
1.
2.
3.
4.
5.
Restore the Licensed Internal Code to the primary partition.
Recover the configuration data for the logical partitions.
Recover the disk configuration for the primary partition.
Install the operating system on the primary partition.
Recover any of the secondary partitions by restoring the Licensed
Internal Code, recovering the disk configuration, and installing the
operating system.
These steps are covered in more detail in the "Recovering Information on
Your System" section in the Backup and Recovery manual, SC41-5304.
• You work with configuration data for logical partitions through a menu in
Dedicated Service Tools. The configuration data for the logical partitions is
never restored from your backup media.
• For secondary partitions that are sharing devices, make sure that you
switched the device to the required logical partition before trying to
perform a recovery.
Problem Management for a Logical Partition
77
• You can view System Reference Code activity for secondary partitions
through the primary partition in Dedicated Service Tools or System Service
Tools. You can also view System Reference Codes activity for a secondary
partition from its own Dedicated Service Tools display or its own System
Service Tools display. You monitor the System Reference Code on the
Display Partition Status screen.
• Look for all configuration data errors that are related to logical partitions in
your Licensed Internal Code log and Product Activity Log. Each logical
partition also has its own logs that you can view. However, some errors on
a secondary partition can only be found in the logs on the primary
partition. Be sure to check both places when possible.
5.7.2 Information on Recovering after System Initialization
Anytime you initialize the disk units on the primary partition, you need to
recover your configuration data. Initialization can occur just prior to
reinstalling the operating system. You need to recover the configuration
because initializing the disk units erases everything on them, including user
data and configuration data.
When you recover your configuration, the system looks for the configuration
data that is stored on the secondary partitions. Then, it copies the
configuration data to the primary partition. When you erase the disk units in a
secondary partition, the configuration data is automatically copied from the
primary partition to that secondary partition once you restart it.
First, you re-install the Licensed Internal Code on the primary partition.
Immediately after the system restarts and before you configure disk units or
recover Auxiliary Storage Pool disks, follow the steps to recover your
configuration as shown in Figure 40 on page 101.
5.7.3 Restoring User Data to a Logical Partition
Restoring user data to a logical partition follows most of the same steps as
restoring data to a nonpartitioned system. However, there are a few
considerations to keep in mind before you begin recovery procedures. See
the "Recovering Information on Your System" section in the Backup and
Recovery manual, SC41-5304, for the specific instructions on data recovery
for your system.
5.7.4 Recovering User Data from a Deleted Logical Partition
In general, you cannot recover user data or system data from a deleted
logical partition. However, if a logical partition is accidentally deleted, it may
78
Slicing the AS/400 with Logical Partitioning: A How to Guide
be possible to manually recreate it without any data loss. Disk units from the
deleted partition must not meet any of the following conditions:
• Moved to another logical partition
• Removed from the system
• Configured into another logical partition or system
To recover the user data and system data, perform the following steps:
1. Create a new logical partition.
2. Add all resources back into the deleted logical partition so that it has the
same configuration as it did before it was deleted.
3. When you finish creating the logical partition, restart the system in normal
B mode.
4. Recover the user data and system data from the disk units.
5.8 Service Provider Issues
Servicing a partitioned system is like servicing multiple separate systems. A
failure in a secondary partition would allow the other partitions to continue
operations. However, there are certain failures that are not partition
dependent that cause the entire system to halt:
•
•
•
•
•
Processor failure
Main storage failure
Service processor (MFIOP)
Power
Any time the primary partition is down, and then all partitions are down!
There are questions that come up when servicing, such as where is my
console, load source, or alternate Initial Program Load device for the
secondary partitions. Since they can physically be located in different frames,
it is important to do the following labeling or have printouts available:
• Rack configuration and partition configuration
• Consoles labeled for each partition
• Label racks if bus-level partitioned
Problem Management for a Logical Partition
79
Figure 31. System Properly Labeled
Concurrent maintenance on a device must be done within the partition to
which the input/output processor is dedicated.
80
Slicing the AS/400 with Logical Partitioning: A How to Guide
Chapter 6. Capacity Planning and Performance
In Chapter 1, “Logical Partitioning Concepts” on page 1, we learned that
logical partitioning can be extremely useful in fulfilling a wide range of
customer needs. The success or failure of an implementation of logical
partitioning in any situation depends largely upon an appropriate analysis of
the situation’s requirements.
6.1 Determining Customer Requirements
The basic questions to be asked before contemplating any kind of solution
are:
• What is the customer trying to achieve?
• How is the customer’s business structured?
• What is the best solution possible for their particular situation?
Consider the following example:
You are approached by a manufacturing company that wants to consolidate
its worldwide operations. Currently, it has five branches in five different
countries. Each branch has its own IT division, that runs the same
applications, and transfers the results of the day’s activities daily to the
machine at headquarters. Here, all results are consolidated, manufacturing
schedules are established, and consolidated reports are compiled. The CEO
has decided that:
• One daily transfer of the day’s activities does not provide him with enough
control of the company’s manufacturing schedules and current financial
situation. He needs instant access to up-to-date information.
• Running five IT divisions is becoming too costly.
• The company needs to consolidate its IT divisions into one at
headquarters.
There are a number of solutions to this unlikely situation. For purposes of
discussion, let us assume that you decide to offer a logical partitioning
solution. You now need to go through some or all of the following steps:
1.
2.
3.
4.
5.
Analyze the workload on each physical machine (maybe using BEST/1).
Determine the amount of interactive and batch resources needed.
Determine languages to be used.
Determine the appropriate communications requirements.
Understand service level agreements and problem or change
management disciplines.
© Copyright IBM Corp. 1999
81
Up to this point, you followed the standard steps towards any solution. From
now on, your focus is narrowed down to logical partitioning issues.
You now need to perform some capacity modelling to determine the batch
and interactive resources required per partition. This will help determine what
type of physical machine is needed. BEST/1 is a useful tool in this respect. In
6.6.2, “BEST/1” on page 99, we go into the finer details of a workable
methodology to determine the resources required using BEST/1. For now, we
look at the bare mechanics of the process.
Decide the machine model, number of partitions, batch or interactive
resources required per partition, number of processors, and amount of main
storage per partition. Now, you are ready to enter the next phase of this
process: Planning for the full hardware configuration. Refer to Chapter 2,
“Planning Considerations” on page 17, for the procedures to follow.
The following sections describe the steps for the actual allocation of
resources to create a multi-partitioned machine for our hypothetical company.
6.2 Commercial Processing Workload (CPW) Concepts
Commercial Processing Workload (CPW) represents a measure of relative
system performance. Different AS/400 models have different CPW ratings.
Traditional AS/400 systems have a single CPW value, which represents the
maximum workload for this model. This CPW value is valid for an interactive
workload, a client/server workload, or a combination of the two workloads.
6.2.1 AS/400 Workload — Technical View
From a technical point of view, the AS/400 workload is divided into two major
sections:
• Interactive workload
The interactive workload or job allows users to see a 5250 display. It can
be a non-programmable terminal (dumb terminal), PC with emulation
software, or an Internet or intranet function with Workstation Gateway. It
usually has a priority level of 20.
• Non-interactive workload
In a non-interactive workload, users do not see a 5250 display. An
example of this workload is a PC program that accesses the AS/400
database using a client/server technique and submitting batch jobs.
AS/400 system jobs are also non-interactive types of work. Most of the
82
Slicing the AS/400 with Logical Partitioning: A How to Guide
non-interactive jobs have a run priority greater than level 20. Some of
them, such as system jobs, have a priority less than level 20
Note: 0 is the highest priority level, and 99 is the lowest priority level.
6.2.2 Different CPW Values
On a server model, there are two CPW values. The larger CPW value shows
the maximum workload for that model if only client/server applications (no
interactive tasks) were running. The smaller number shows the maximum
workload for that model if only interactive applications were running. These
two values are not additive. This means that interactive CPW and
client/server CPW cannot be added together since interactive CPW is part of
the client/server CPW value. Interactive tasks reduce the client/server
processing capability of the system.
Custom server models act similar to server models. But, custom servers
follow a different algorithm. Therefore, the "knee" of the curve for workload
interaction is at a different point.
The AS/400 server 7xx models (announced February 1999) and the new
Model 170 servers (announced September 1998) also have two CPW values.
The higher value shows the relative processor capacity. Therefore, there is
no need to differentiate between constrained and unconstrained CPW values.
The smaller CPW value represents the maximum interactive workload for the
specific model. If you add an interactive workload to your Model 170, the
remaining client/server capacity available is calculated as:
CPW (C/S batch) = CPW (processor) - CPW (interactive)
In performing capacity planning for logical partitions, it is important to know
what the specific algorithm for that AS/400 model is. On the new AS/400 7xx
models, you can create a logical AS/400 partition that:
•
•
•
•
Suits your own specific needs
Works like a traditional AS/400 system model
Works like a custom server model
Works like an AS/400 server model
6.2.3 AS/400 System Models
For traditional AS/400 system models, the full CPU processor capabilities are
available to either standalone interactive, stand-alone non-interactive, or
mixed-mode work environments. The impact between is are controlled by the
job Run Priority, Time Slice End, and Storage Pool parameters.
Capacity Planning and Performance
83
Figure 32 shows how system performance is on an AS/400 Model 600 if you
mix interactive and client/server (batch) workloads.
Figure 32. Interactive CPW Utilization — Model 600
Table 3 offers another way to view performance based on specific interactive
percentages.
Table 3. AS/400 Model 600 — Feature Processor #2136
Interactive Percentage
0%
33%
66%
100%
CPW interactive
0.0
24.1
48.2
73
CPW client/server
73
48.7
24.3
0
6.2.4 New AS/400e server 7XX
Only for the new February 1999 7XX processors and the September 1998
170 processors, the system will manage the interactive work differently than
on other servers. In an interactive-only environment, the interactive jobs
(5250 green screen jobs) will get the rated CPW (or slightly higher CPW
rating). However, the CPU may not be 100% utilized. You will still find that the
non-interactive-only environment gets the full non-interactive CPW, now
called Processor CPW rating. In an environment of both interactive and
non-interactive work, the interactive work will be permitted to reach its CPW
rating, but not any higher.
The new 7XX and 170 models have interactive and processor CPW ratings.
The 170 models have a fixed interactive rating. For the new 7XX models, you
84
Slicing the AS/400 with Logical Partitioning: A How to Guide
can choose the interactive CPW you need for your workload. In a mixed
interactive and non-interactive workload environment, you can size the
system based on subtracting interactive CPW from the full non-interactive processor CPW rating.
For example, a Model 730-2068 has a processor CPW rating of 2,890 and an
interactive CPW rating (with interactive feature 1511) of 2,000. When the
interactive jobs use 2,000 CPW (actually the system manages this as a
percentage of CPU utilization), no more interactive job throughput can be
achieved. However, at that level, there is CPU left for non-interactive jobs to
achieve 890 CPW amount of work.
Figure 33 shows how system performance is on an AS/400 Model 730 if you
mix interactive and client/server (batch) workloads.
Figure 33. Interactive CPW Utilization — Model 730
Table 4 offers another way to view performance based on specific interactive
percentages.
Table 4. AS/400 Model 730 — Feature Processor #2068
Interactive Percentage
0%
33%
66%
100%
CPW interactive
0.0
660
1320
2000
CPW client/server
2890
2230
1570
890
Capacity Planning and Performance
85
6.2.5 AS/400 Server Models
For server models, the full CPU processor capabilities are available only to
non-interactive work environments. The non-interactive CPW rating is
achieved in a non-interactive only job environment at 100% CPU utilization.
For the interactive CPW rating, 5250 work consumes 100% CPU utilization,
but cannot make full use of the maximum CPU power.
At approximately one third of the interactive CPW rating, the interactive work
begins to degrade any concurrent non-interactive work. The impact
accelerates dramatically as interactive work grows above this one-third CPW
value.
Figure 34 shows how system performance is on an AS/400 Model S30 with
processor feature 2259 if you mix interactive and client/server (batch)
workloads.
Figure 34. Interactive CPW Utilization — Model S30 Processor #2259
Table 5 offers another way to view performance based on specific interactive
percentages.
Table 5. AS/400 Model S30 — Feature Processor #2259
86
Interactive Percentage
0%
33%
66%
100%
CPW interactive
0.0
21.1
42.2
64.0
CPW client/server
998.6
931.9
466.0
0.0
Slicing the AS/400 with Logical Partitioning: A How to Guide
6.2.6 AS/400 Custom Server Models
For custom mixed-mode server systems, the full CPU processor capabilities
are available only to non-interactive work environments. The non-interactive
CPW rating is achieved in a non-interactive only job environment at 100%
CPU utilization. For the interactive CPW rating, 5250 work consumes 100%
CPU utilization, but cannot make full use of the raw CPU power.
The relatively higher interactive CPW rating on Custom Mixed Mode servers
is achieved by the system permitting interactive work to grow until it is at six
sevenths of the interactive CPW rating. Above this six-sevenths CPW rating,
the interactive work impact on non-interactive work dramatically increases.
Figure 35 shows how system performance will be on an AS/400 Model S20 if
you mix interactive and client/server (batch) workloads.
Figure 35. Interactive CPW Utilization — Model S20
Table 6 offers another way to view performance based on specific interactive
percentages.
Table 6. AS/400 Model S20 — Feature Processor #2170
Interactive Percentage
0%
33%
66%
100%
CPW interactive
0.0
16.4
32.8
49.7
CPW client/server
464.3
447.4
430.9
0.0
Capacity Planning and Performance
87
6.3 Determining Batch CPW
Non-interactive workload includes the typical batch job as well as the AS/400
client/server workload. It is triggered by the Submit Job command, spooling,
message queue, and data queue entries.
Typically batch functions include:
• Networking
– APPC
– TCP/IP
• Database
– SQL
– DDM
– Remote SQL
– ODBC
– File transfer
• File serving
– Shared folders
• Client/server applications
– Progress/400
– SYNON CSG
– SQL Windows
– ENVY/400
– And so on
• Traditional batch
– Development
– Data warehouse
– Internet/intranet
– Domino/Notes
A non-interactive transaction includes:
•
•
•
•
•
•
88
Job run time
Time to complete printing a spooled file
Time to complete a set of related jobs
Time to process "n" message queue or data queue entries
The number of records processed per hour
Time to send or receive a file or files between a server and its attached
client (or clients)
Slicing the AS/400 with Logical Partitioning: A How to Guide
The performance objectives should show the amount of work performed and
the amount of resources used in a specific amount of elapsed time. The
amount of resources to perform this work is important as well since it defines
the amount of CPU needed, the number of DASD arms, or the amount of
DASD space needed.
Knowing the number of records processed per unit of time, the number of
disk I/O operations per second, or the number of bytes (characters)
transferred by these non-interactive jobs is, in many cases, the only way to
set realistic expectations or evaluate degradation or improvement in batch
run times.
6.3.1 BATCH400 — Enabling the AS/400 Batch Window
BATCH400 is a new AS/400 tool to help analyze and predict batch job
performance on the AS/400 system. BATCH400 addresses the often asked
question: What can I do to my system to meet my overnight batch runtime
requirements (also known as the Batch Window)?
The BATCH400 tool creates a "model" from AS/400 performance data. This
model resides in a file in the target library. The tool can be asked to analyze
the model and provide results for various "what-if" conditions. Individual batch
job runtime and overall batch window runtimes are reported by this tool.
BATCH400 allows you to "view" the batch window in a form that shows job
start and stop times and dependencies. This should allow you to locate times
in the batch window when more efficient job scheduling can improve total
system throughput. BATCH400 can also help predict improvements in
throughput that results from such hardware upgrades as CPU and DISK.
Note: Batch400 is an IBM internal-use-only tool.
6.3.2 Batch Capacity Planning for LPAR
Before you start capacity planning, be sure you have the right hardware and
software requirements fulfilled for your system. Only OS/400 Version 4
Release 4 (V4R4) supports logical partitioning. V4R4 is the oldest release
supported in any partition.
Capacity Planning and Performance
89
Logical partitioning is only supported on the AS/400 systems shown in Table
7.
Table 7. Systems Supporting LPAR — Batch CPW
Model
2-Way
CPW
Batch
620
#2182
464.3
640
#2238
583.3
4-Way
CPW
Batch
#2239
998.6
650
720
#2063
810
#2064
1,600
730
#2066
1,050
#2067
2,000
740
8-Way
CPW
Batch
12-Way
CPW
Batch
#2240
1,794
#2243
2,340
#2188
3,660
#2189
4,550
#2068
2,890
#2069
3,660
#2070
4,550
S20
#2165
464.3
#2166
759.0
S20
ISV
#2170
464.3
#2177
759.0
#2178
759.0
#2259
998.6
#2260
1,794
#2320
998.6
#2321
1,794
#2322
1,794
#2256
1,794
#2261
2,340
#2207
3,660
#2208
4,550
#2340
3,660
#2341
4,550
S30
S30
ISV
#2258
583.3
S40
S40
ISV
To determine the right batch CPW value for your partitioned system, you have
to know the CPW value for that system. In Table 7, you can find the CPW
value available for the eligible systems. On an N-way system, each processor
has a specific CPW value. Divide the total CPW value by the number of
processors to calculate the individual CPW per processor. If the value for a
single processor meets your requirements for batch workload, you can setup
a logical partition with one processor only. If you need more CPW for your
batch workload, you have to add one or more of the available processors to
that specific partition.
90
Slicing the AS/400 with Logical Partitioning: A How to Guide
6.3.2.1 Example: Assigning CPW for a Batch
Before you start assigning CPW value for logical partitions, we strongly
recommend that you do capacity planning. In this example (see Table 8),
capacity planning showed a CPW value of 550 required for your batch
environment.
Table 8. Assigning CPW for a Batch Environment
AS/400 Model
9406-720 #2064
No. of Processors Available
4
CPW Available
Required
Batch CPW
1,600
550
First, you need to calculate the CPW value for each processor:
CPW (single processor) = CPW (batch) / number of processors
CPW (single processor) = 1,600 / 4 = 400 CPW
The maximum CPW value you can get in a partition with one processor only
is 400 CPW. In this example (see Figure 36), you need 550 CPW for your
batch workload. You need at least two processors for your batch workload.
Only whole processors can be assigned to a logical partition. During setup for
a logical partition, you are also asked for the amount of main storage for that
partition. To determine the required amount of main storage, you have to
perform capacity planning using BEST/1.
Figure 36. Assigning CPW for Batch
Capacity Planning and Performance
91
6.4 Determining Partition Processing Resources
In the process of setting up your partitions, you need to ensure that enough
resources are available to provide the required machine performance. The
following discussion attempts to show the pure mechanics of working out
your logical partitioning configuration. This may not lead to a feasible
configuration. Therefore, at a later time, we need to consider the pitfalls to
avoid and refinements to apply in order to create a logically partitioned
machine that can actually be physically implemented.
In the planning stages for your machine, you decide the overall scope of most
of these parameters. Now, we drill down to the process of logically
partitioning the machine by allocating the appropriate resources. Figure 37
shows the actual screen from which you will be working.
Create New Partition
System:
SYSTEMA
Complete blanks, press Enter.
Partition identifier . . . . . . . . . . . . . . 0
Partition name . . . . . . . . . . . . . . . . .
Number of available system processors . . . . : 8
Number of partition processors . . . . . . . . . 1
Minimum / maximum number of processors . . . . . 1
/ 2
Size of available system main storage (MB) . . : 24576
Size of partition main storage (MB) . . . . . . 2048
Minimum / maximum size of main storage (MB) . . 2048
Interactive performance available . . . . . . : 100 %
Partition interactive performance . . . . . . . 12 %
Minimum / maximum interactive performance . . . 12 /12
Connect to system inter-partition OptiConnect .
F3=Exit
F11=Display partition processing configuration
/
%
1=Yes, 2=No
F9=Exclude limits
F12=Cancel
Figure 37. Creating Logical Partitions
The three resources we need to consider now are processors, main storage,
and interactive performance. We examine these items one at a time in the
following sections.
92
Slicing the AS/400 with Logical Partitioning: A How to Guide
6.4.1 Determining the Number of Processors
In your planning, before ordering your machine, you decide on:
• The number of partitions required (in our case, 5)
• The CPW requirements for each partition.
These are obtained by modelling through BEST/1. For example, one
environment produced a Model 720 with a processor CPW rating of 810.
Refer to 6.2, “Commercial Processing Workload (CPW) Concepts” on
page 82, for more details. You know that the partition to contain that
environment needs this CPW rating. As an example, Table 9 provides
details of the new 7XX models.
Table 9. CPW Ratings for 7XX Models
Model
Initial
Proc.
Code
# of
Proc.
#
Proc.
CPW
Inter.
Base
Additional Interactive CPW
1501
1502
1503
1504
1505
720
2061
1
240
35
70
120
720
2062
1
420
35
70
120
240
720
2063
2
810
35
120
240
560
720
2064
4
1,600
35
120
240
560
1,050
1507
1508
1509
1510
1511
730
2065
1
560
70
120
240
560
730
2066
2
1050
70
120
240
560
1,050
730
2067
4
2000
70
240
560
1,050
2,000
730
2068
8
2890
70
240
560
1,050
2,000
1510
1511
1512
1513
740
2069
8
3,660
120
1,050
2,000
3,660
4,550
740
2070
12
4,550
120
1,050
2,000
3,660
4,550
Capacity Planning and Performance
93
The complete capacity planning results for our hypothetical situation are
summarized in Table 10.
Table 10. Capacity Planning Results
Environment
1
Environment
2
Environment
3
Environment
4
Environment
5
Model: 720
Feature: 2061
Prc.CPW: 240
Int. CPW :120
Model: 720
Feature: 2063
Prc.CPW: 810
Int. CPW :240
Model: 720
Feature: 2063
Prc.CPW: 810
Int. CPW :240
Model: 720
Feature: 2061
Prc.CPW: 240
Int. CPW :120
Model: 720
Feature: 2062
Prc.CPW: 420
Int. CPW :240
A possible mapping of the above into a logical partitioning context would
appear as shown in Table 11.
Table 11. First Cut into Logical Partitions
Partition 0
Partition 1
Partition 2
Partition 3
Partition 4
Prc.CPW: 240
Int. CPW :120
Prc.CPW: 810
Int. CPW :240
Prc.CPW: 810
Int. CPW :240
Prc.CPW: 240
Int. CPW :120
Prc.CPW: 420
Int. CPW :240
This would have yielded a physical machine with a minimum of 2,520
processor CPW and 960 Interactive CPW. The nearest available machine
matching these parameters is an 8-Way Model 730-2068 rated at 2,890
processor CPW and 1,050 Interactive CPW. With such a machine, the
CPW per processor is 361 (2,890 divided by 8).
You can now work out the number of processors per partition as shown in
Table 12.
Table 12. Working Out the Number of Processors
Partition 0
Partition 1
Partition 2
Partition 3
Partition 4
Prc.CPW: 240
Int. CPW :120
Prc.CPW: 810
Int. CPW :240
Prc.CPW: 810
Int. CPW :240
Prc.CPW: 240
Int. CPW :120
Prc.CPW: 420
Int. CPW :240
1 Processor
361 CPW
3 Processors
1,083 CPW
2 Processors
722 CPW
1 Processor
361 CPW
1 Processor
361 CPW
Notice that the processing power of a partition is tightly coupled with the
number of processors it contains. For example, partition 1 needs three
processors to fulfill its CPW quota. Only one processor is required for the
primary partition (partition 0).
94
Slicing the AS/400 with Logical Partitioning: A How to Guide
Looking at partition 2, we notice that our capacity plan specified 810
processor CPW, but we managed only 722 CPW. In a real situation, this
would alert us to the fact that we need a physical processor with more CPW
and need to go back to the drawing board. However, we will ignore this for
this exercise.
The logical partitioning function allows you to specify a minimum and
maximum range for the number of processors for each partition. This
capability can be very useful in managing changes to a partition workload.
For example, if, for some reason, there is a reduction in the workload in
partition 1 and you find that it now needs only 600 CPW, you have the
flexibility of reducing the number of processors, as long as you are within
your defined range. The removed processors become available for an
addition to any other partition needing more power, with minimum disruption.
Notes:
• Changes to processor resources beyond the defined range require a full
system restart, a major disruption.
• The absolute minimum is one processor per partition. These minimums
can be specified only if the maximum limit for the partition is the same.
6.4.2 Determining Main Storage
In planning user requirements, you would use a tool like BEST/1 to determine
the amount of main storage needed. Table 13 shows how this would work.
Table 13. Working Out Memory Allocation
Partition 0
Partition 1
Partition 2
Partition 3
Partition 4
Prc.CPW: 240
Int. CPW :120
Prc.CPW: 810
Int. CPW :240
Prc.CPW: 810
Int. CPW :240
Prc.CPW: 240
Int. CPW :120
Prc.CPW: 420
Int. CPW :240
1 Processor
361 CPW
3 Processors
1,083 CPW
2 Processors
722 CPW
1 Processor
361 CPW
1 Processor
361 CPW
2,048 MB
7,068 MB
4,096 MB
1,024 MB
3,192 MB
You decide to add 35% more memory because you now have a good
understanding of the customer’s workload. This yields a Model 730 with
24,576 MB of main storage.
Using the screen pictured in Figure 37 on page 92, you can now assign main
storage to each of your partitions. Again, specify the minimum and maximum
range to allow some flexibility for on-the-fly main storage changes with
minimum disruption.
Capacity Planning and Performance
95
Notes:
• Changes to main storage resources beyond the defined range require a
full system restart, a major disruption.
• The absolute minimum main storage is 256 MB for the primary partition
and 64 MB for the secondary partition.
6.4.3 Determining Interactive Performance
One of the most important conclusions you would reach in the initial analysis
of your customer’s environment is the type of workload in each environment,
namely batch (client/server) or interactive (green screen). Your assessment
would directly affect your choice of processor feature.
Refer to Table 10 on page 94 for the complete results of your capacity
planning exercise.
You are now ready to assign interactive performance to each of your
partitions.
Table 14. Working Out Interactive Performance
Partition 0
Partition 1
Partition 2
Partition 3
Partition 4
Prc.CPW: 240
Int. CPW :120
Prc.CPW: 810
Int. CPW :240
Prc.CPW: 810
Int. CPW :240
Prc.CPW: 240
Int. CPW :120
Prc.CPW: 420
Int. CPW :240
1 Processor
361 CPW
3 Processors
1,083 CPW
2 Processors
722 CPW
1 Processor
361 CPW
1 Processor
361 CPW
2,048 MB
7,068 MB
4,096 MB
1,024 MB
3,192 MB
Actual: 12%
Min: 1%
Max: 12%
Actual: 37%
Min: 3%
Max: 37%
Actual: 24%
Min: 2%
Max:24%
Actual: 12%
Min: 1%
Max: 12%
Actual: 12%
Min: 1%
Max: 12%
From a performance perspective, it is important to note that with the new 7XX
models, the published interactive CPW is fully usable to 100% without
incurring any performance penalty. This is different from prior models (for
example, Model S30). In the earlier models, a performance penalty is
incurred when interactive CPU usage reaches either one-third or
six-sevenths of the published interactive CPW, depending on the model (Sxx
or the newer mixed-mode servers).
When dealing with non-7XX server models, interactive performance
calculations must take into account the fact that there is a drastic
performance degradation if the CPU percentage of interactive usage exceeds
the limits previously mentioned.
96
Slicing the AS/400 with Logical Partitioning: A How to Guide
Note:
• Changes to interactive resources beyond the defined range require a full
system restart, a major disruption.
• The absolute minimum interactive performance is 1% for each processor.
This means, for example, that a three-processor partition will have a
minimum interactive performance of 3% on a model 7XX.
• Notice that the maximum interactive performance that you can obtain
depends on the number of processors in a partition. For example, in the
first partition in Table 14 on page 96, the maximum interactive
performance is 12%. This is the percentage of the interactive CPW for one
processor out of eight. In the same way, in the second partition, which has
three processors, the maximum interactive performance possible is 37%
(three-eighths of the total interactive performance).
Refer to Appendix E, “Interactive and Batch CPW Determination” on page
189, for a more detailed scenario of the process.
6.5 OptiConnect/400 Performance Considerations
OptiConnect/400 allows applications written to access databases locally, to
access them remotely with simple changes to file descriptions and no
changes to applications. It provides an optimized Distributed Data
Management (DDM) solution that is capable of efficient, low-latency, high
bandwidth communication between AS/400 logical partitions.
Before V4R4 and logical partitioning, OptiConnect/400 utilized a fiber-optic
connected, shared I/O bus between AS/400 systems. With V4R4,
OptiConnect/400 can use either the special hardware I/O processors or the
virtual OptiConnect on a single N-way system with logical partitions.
The following conclusions are derived from laboratory benchmarks and from
experience gained by working with IBM OptiConnect/400 customers:
• Each time that a put or get is issued, OptiConnect/400 uses the shared I/O
bus to exchange data with the remote system. DDM uses a
communication link for accessing data on a remote system. In fact, this is
different if you use virtual OptiConnect on logical partitions. Virtual
OptiConnect does not send out any data to the shared bus, but uses a
copy function of the memory to transport data from one partition to the
other, using TCP/IP protocols.
• OptiConnect/400 is a more efficient way than traditional DDM using APPC
controllers to access data on another AS/400 partition.
Capacity Planning and Performance
97
• At higher levels of throughput, the system overhead for OptiConnect/400
increases in a non-linear fashion. This is due to multiprocessing
contention when CPU utilization is higher. For this reason, we recommend
that CPU utilization should not exceed 90% on either the application
machine or database server system.
• There is an overhead (called the client/server factor ) associated with using
OptiConnect/400 both on the application and database machine. This
overhead varies and depends on such factors as:
– Percent of database activity
– Number of logical I/Os per transaction
– Usage of journal and commitment control
– Number of database opens and closes
• When analyzing batch workloads, pay more attention to logical database
I/Os, blocking, and pool size. Performance characteristics may be different
for batch versus interactive use of OptiConnect/400. In particular, batch
jobs are often database intensive, and therefore, accumulate CPU and I/O
delays for each logical I/O that is requested from the other system. This
can add significant CPU utilization and elapsed time to batch runs.
• A detailed analysis to determine where to place the data files is critical in
batch workloads or read-only files.
• In most distributed environments, including OptiConnect/400, there is a
penalty for remote access. Obviously this penalty is most severe when
100% of accesses are remote and may become negligible for lower
percentages (for example, 10%). Therefore, an optimum environment
should minimize remote access by distributing users or data across
systems in a manner to maximize local access. Carefully consider the
replication of read-only files to application machines, the distribution of
data across multiple database servers, and the location of batch work on
application machines or database servers when defining an
OptiConnect/400 environment to optimize total system capacity.
6.6 Available Tools
There are no tools to provide consolidated performance and capacity
planning across a multi-partitioned system. Any result produced by such a
tool would be meaningless. Logical partitions are independent systems and,
as such, they must be managed as if they were separate, physical systems. It
follows that all standard tools currently available to analyze system
performance or performing capacity planning functions on physical machines
can also be used most naturally on any single partition independently.
98
Slicing the AS/400 with Logical Partitioning: A How to Guide
The same methodology applies. For each partition, you need to:
•
•
•
•
•
•
•
Establish performance objectives
Develop a good strategy to approach the problem
Measure current performance
Perform data analysis
Perform trend analysis
Perform problem analysis
Setup benchmark for progress evaluation
In the next few sections, we explain some of the tools available and go into
detail for those such as Management Central, which have been extended to
facilitate the task of managing performance on multi-partitioned and clustered
systems.
6.6.1 Performance Tools
The role of Performance Tools has not changed. In the initial planning stages
to implement a logical partitioning configuration, Performance Tools are used
for performance data collection, performance data analysis, and problem
analysis. These steps are still done on each single partition in the same way
as they are done on any non-partitioned system.
6.6.2 BEST/1
BEST/1 is the tool of choice to help determine hardware requirements as a
preliminary to creating a logically partitionable machine.
After you work out the broad lines of your configuration, you may not be able
to find a physical machine that matches your preferred configuration. For
example, you may have decided that you need a 7XX model with a
three-processor logical partition. This is not a valid physical machine, but you
still need to measure how such a logical machine will handle the workload
that you are going assign to it.
The solution to this dilemma is to create a new customized processor feature
code using BEST/1. The resulting configuration can then be used for
standard modelling. The next section explains the methodology to do it.
6.7 Creating a New Hardware Feature in BEST/1
If you created your own individual logical partition on an AS/400 model 7XX,
you cannot find this model information in the hardware file in BEST/1. In
BEST/1, you will find all of the 7xx models and related feature codes you can
order, for example, an AS/400 Model 740 with a processor feature code of
Capacity Planning and Performance
99
2E6D. However, if you create a logical partition on a Model 720 with, for
example, two processors that each have 400 CPW and your interactive
performance will be at 46%, you have to create your own specific hardware
feature in BEST/1. This section guides you through the setup of a new
hardware feature in BEST/1.
6.7.1 Going to the Partition Processing Configuration
To go to the partition processing configuration, you have to start System
Service Tools (SST). To do this, complete these steps:
1. Enter the command:
STRSST
2. Select option 5, Work with system partitions (Figure 38).
System Service Tools (SST)
Select one of the following:
1.
2.
3.
4.
5.
Start a service tool
Work with active service tools
Work with disk units
Work with diskette data recovery
Work with system partitions
Selection
5
F3=Exit
F10=Command entry
F12=Cancel
Figure 38. Going to the Partition Processing Configuration (Part 1 of 4)
3. On the Work with System Partitions screen, select option 1 (Figure 39 on
page 101).
100
Slicing the AS/400 with Logical Partitioning: A How to Guide
Work with System Partitions
System:
LPAR2
Attention: Incorrect use of this utility can cause damage
to data in this system. See service documentation.
Number of partitions . . . . . . . : 2
Partition manager release . . . . . : V4R4M0
Partition identifier . . . . . . . : 1
Partition name . . . . . . . . . . : LPAR2
Select one of the following:
1. Display partition information
2. Work with partition status
4. Recover configuration data
Selection
1
F3=Exit
F12=Cancel
Figure 39. Going to the Partition Processing Configuration (Part 2 of 4)
4. On the Display Partition Information screen, select option 2 (Figure 40).
Display Partition Information
System:
LPAR2
Select one of the following:
1.
2.
3.
4.
Display
Display
Display
Display
partition
partition
allocated
available
status
processing configuration
I/O resources
I/O resources
Selection
2
F3=Exit
F12=Cancel
Figure 40. Going to the Partition Processing Configuration (Part 3 of 4)
Capacity Planning and Performance
101
Display Partition Processing Configuration
System:
LPAR2
Number of system processors . . . . . . . . . . . : 4
Size of system main storage (MB) . . . . . . . . : 8192
Partition
Identifier Name
1
PAR2
Total
Main Storage Interactive Inter-partition
Processors Size (MB)
Percentage OptiConnect
2
4096
46
Yes
Only this partition is shown, use primary to see all partitions.
F3=Exit F11=Display allocated I/O resources F12=Cancel
Figure 41. Going to the Partition Processing Configuration (Part 4 of 4)
You now have all the information you need to create a new hardware feature
in BEST/1.
In our example, we have a base system with four processors and 8,192 MB of
main storage. Actually, this AS/400 system is a Model 720 with processor
feature code 208D. In Figure 42 on page 104, you can see the relative
performance rating of this model. The processor performance value is
403.69. The value for the interactive performance is 219.41.
The performance values in BEST/1 are still based on Model B10, which has a
relative performance rating of 1.0. The published CPW values for all of the
AS/400 systems cannot be found in BEST/1. We do not need them for our
new hardware feature. We need the performance values found in Figure 42
for the calculation of the new hardware feature in BEST/1.
6.7.2 Calculating New Performance Ratings for Logical Partitions
To calculate the new performance ratings, follow the process described here:
The processor performance rating is 403.69. The base system has four
processors. Determine the performance rating by using this equation:
403.69 / 4 = 100.92
This produces a performance rating of 100.9 for each processor.
102
Slicing the AS/400 with Logical Partitioning: A How to Guide
Now calculate the relative performance rating for the batch:
100.92 * 2 = 201.84
This produces a relative performance rating for batch of 201.84. The
interactive performance rating is 219.41. The interactive percentage in logical
partition is 46%. 46% of 219.41 is 100.93.
The maximum available relative performance rating is 201.84. Our interactive
performance rating is 100.93. These are the values that we need when we
create our new hardware feature in BEST/1.
6.7.3 Creating the New Hardware Feature in BEST/1
To create a new hardware feature member in the BEST/1 hardware file, follow
these steps:
1. Type STRBEST and press Enter.
2. Press Enter on the disclaimer screen.
3. On the BEST/1 for the AS/400 screen, type 60 (More BEST/1 Options) and
press Enter.
4. On the More BEST/1 Options screen, type 10 (Hardware characteristics
menu) and press Enter.
5. On the Hardware Characteristics screen, type 1 (Work with CPU models)
and press Enter.
6. On the list of CPU models, find one with the same CPU value (in our
example, 720) and a model value that comes close to the one noted (in
our example, 208D).
7. On the Work with CPU Models screen, type 3 (copy) next to the model that
you selected in the previous step (Figure 42 on page 104) and press
Enter.
Capacity Planning and Performance
103
Work with CPU Models
Type option, press Enter.
1=Create 2=Change 3=Copy
Opt CPU Model
3
720
720
720
720
720
720
436
436
436
170
170
S20
S20
F3=Exit
207E
207F
208A
208B
208C
208D
2102
2104
2106
2159
2160
2161
2163
4=Delete
7=Rename
System Archi- --Relative Perf--- Number of Max Stor
Unit tecture
Normal
Server Processors
(MB)
9406
9406
9406
9406
9406
9406
9402
9402
9402
9406
9406
9406
9406
*RISC
*RISC
*RISC
*RISC
*RISC
*RISC
*RISC
*RISC
*RISC
*RISC
*RISC
*RISC
*RISC
104.27
9.05
30.77
61.55
109.70
219.41
4.80
6.10
8.70
3.60
5.10
6.50
8.00
199.86
403.69
403.69
403.69
403.69
403.69
22.50
29.30
29.30
57.70
F9=Display equivalent BEST/1 and IBM CPU models
2
4
4
4
4
4
1
1
1
1
1
1
1
8192
8192
8192
8192
8192
8192
224
224
256
832
832
2048
2048
More...
F12=Cancel
Figure 42. Work with CPU Models — BEST/1
8. On the Copy CPU Model screen, type in the new name (in our example,
LPAR) and press Enter. You will see a 208D is copied at the bottom of the
screen (Figure 43).
Copy CPU Model
To copy, type New Name, press Enter.
Current name . . . . . . . . . . :
208D
New name . . . . . . . . . . . . .
LPAR
F3=Exit F12=Cancel
BEST/1 CPU model 720 208D refers to IBM CPU model 720 2064-1505
Figure 43. Copy CPU Model — BEST/1
104
Slicing the AS/400 with Logical Partitioning: A How to Guide
9. Edit the new model you created by typing a 2 next to it on the Work with
CPU Models screen (Figure 44). Press Enter.
Change CPU Model
CPU model . . . . . . . . . . . . . . . . :
Min/Max storage size (MB) . . . . . . . . :
Type information, press Enter.
System unit . . . . . . . . . . .
Architecture . . . . . . . . . .
Relative performance (B10 = 1.0):
Normal . . . . . . . . . . . .
Server . . . . . . . . . . . .
Number of processors . . . . . .
Currently available . . . . . . .
Family . . . . . . . . . . . . .
Upgrade to family . . . . . . . .
LPAR
256
. .
. .
9406
*RISC
.
.
.
.
.
.
100.93
201.84
2
Y
*POWERAS
*NONE
.
.
.
.
.
.
Disk IOPs . . . . . . . . . . . . . . . . .
Multifunction IOPs . . . . . . . . . . . . .
8192
9402, 9404, 9406
*CISC, *RISC
(Blank if not Server)
Y=Yes, N=No
Name
*NONE, name
Minimum
0
1
Maximum
22
17
F3=Exit F6=Specify storage sizes
F9=Specify connections to disk IOPs
F11=Specify connections to disk drives F12=Cancel F24=More keys
BEST/1 CPU model 720 LPAR refers to IBM CPU model 720 2064-1505
Figure 44. Change CPU Model — BEST/1
10.On the Change CPU Model screen, update the performance values to the
ones we calculated before. Change the number of processors to 2, and
change the value Upgrade to family to *NONE. Then, press Enter.
11.From the Work with CPU Models screen, press Enter.
12.From the Hardware Characteristics screen, press F15 (Save current
hardware characteristics).
13.On the Save Hardware Characteristics screen (Figure 45 on page 106),
enter a member name, (in our example, NEWLPAR). Leave the library name
as QGPL. Press Enter.
Capacity Planning and Performance
105
Save Hardware Characteristics
Change values if desired, press Enter.
Member . . . . . . . . . . . .
Library . . . . . . . . . .
NEWLPAR
QGPL
Name
Name
Text . . . . . . . . . . . . .
New logical partition on 9406-720
Replace . . . . . . . . . . .
N
Y=Yes, N=No
F12=Cancel
Figure 45. Save Hardware Characteristics — BEST/1
14.From the Hardware Characteristics screen, press Enter.
15.From the More BEST/1 Options screen, press Enter.
16.From the BEST/1 for AS/400 screen, press F3 .
You should be back at the command line.
17.Enter the following command:
WRKMBRPDM QGPL/QACYHRWS
The screen shown in Figure 46 on page 107 appears.
106
Slicing the AS/400 with Logical Partitioning: A How to Guide
Work with Members Using PDM
File . . . . . .
Library . . . .
QACYHRWS
QGPL
Position to . . . . .
Type options, press Enter.
3=Copy
4=Delete
9=Save
13=Change text
5=Display 7=Rename
18=Change using DFU
Opt Member
3
NEWLPAR
Text
New logical partition on 9406-720
Date
05/04/99
RCHASM25
8=Display description
25=Find string ...
Bottom
Parameters or command
===>
F3=Exit
F4=Prompt
F9=Retrieve F10=Command entry
F5=Refresh
F6=Create
F23=More options F24=More keys
Figure 46. Work with Members Using PDM
18.On the Work with Members Using PDM screen, type a 3 next to the
NEWLPAR member. Press Enter. The screen shown in Figure 47 appears.
Copy Members
From file . . . . . . . :
From library . . . . :
QACYHRWS
QGPL
Type the file name and library name to receive the copied members.
To file . . . . . . . .
To library . . . . .
QACYHRWS
QPFR
Name, F4 for list
To rename copied member, type New Name, press Enter.
Member
NEWLPAR
New Name
NEWLPAR
Bottom
F3=Exit
F4=Prompt
F19=Submit to batch
F5=Refresh
F12=Cancel
Figure 47. Copy Members
19.On the Copy Members screen, type QPFR in the To Library parameter.
Press Enter.
Capacity Planning and Performance
107
20.On the Work with Members Using PDM screen, press Enter to go to the
command line.
Using these steps, you created a new hardware model in BEST/1. If you are
now sizing a workload or modelling with performance data for the logical
partition, you have the exact hardware characteristics of the logical partition
in BEST/1.
6.7.4 Using the New Hardware Feature
If you started a performance collection using Collection Services function of
Management Central and you created the database files for the Performance
Tools on the AS/400, you can create a BEST/1 model from this performance
data. This performance data will have all of the right information about the
hardware of the partition on which you collected the data. The only values
that are not included are the relative performance ratings for this partition. If
you are going to use the collected data for analyzing and modeling using
BEST/1, you have to go to the configuration menu. Using option 1, you can
change your CPU and other resource values. Figure 48 shows how to specify
the created hardware feature from the Change CPU and other Resource
Values display in BEST/1.
Change CPU and Other Resource Values
Model/Text:
LPAR2
Test TCP/IP FTP on RCHASM25
Type choices, press Enter.
System unit . . . .
CPU Model . . . . .
System storage (MB)
Release level . . .
F3=Exit
F4=Prompt
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
9406
LPAR
4096
V4R4M0
F12=Cancel
Figure 48. Change CPU and Other Resource Values
108
Slicing the AS/400 with Logical Partitioning: A How to Guide
9402, 9404, 9406
F4 for list
F4 for list
F4 for list
If you press F4 and your previously created hardware feature is not shown in
the available hardware features, go to the Hardware Characteristics menu
and press F14. Your created model will appear in the list. Figure 49 shows the
Hardware Characteristics menu.
Hardware Characteristics
Select one of the following:
1.
2.
3.
4.
5.
6.
7.
Work
Work
Work
Work
Work
Work
Work
with
with
with
with
with
with
with
CPU models
Disk IOP features
Disk Drive features
Communications IOP features
Communications line speeds
Multifunction IOP features
IOA features
Selection or command
===>
F3=Exit F4=Prompt F9=Retrieve F12=Cancel
F14=Replace with saved hardware characteristics F24=Morekeys
Figure 49. Hardware Characteristics Menu
6.7.5 Performance Management/400
Performance Management/400 (PM/400) offers a broad array of benefits. It is
a systems management tool that ensures that you get the most out of your
system by regularly monitoring performance and growth. These
measurements allow you, your IBM Business Partner, or IBM to more easily
plan future system growth and identify potential resource constraints.
Measuring your system’s performance helps you make decisions that affect
your budget and human resources plans.
PM/400 is automated, self-managing, and easy to use. PM/400 gathers the
non-proprietary performance and capacity data from the AS/400 system and
automatically sends it to IBM. In return, you receive this data in a series of
reports and graphs that show your system’s growth and performance.
PM/400 is already installed on your systems. It comes with the operating
system, so it is installed in all your partitions. To send data, it uses the same
resources as the Electronic Customer Support line. The IBM "factory" that
Capacity Planning and Performance
109
produces the reports is not set up to handle data from several partitions in
this release, V4R4, since it is driven by a serial number.
Note: All partitions on a machine have the same serial number.
Only data from the primary partition is handled unless special arrangements
are made with IBM. Since this is a restriction in the "factory" and not in the
PM/400 code on your system, it may be subject to change. All features in
PM/400 on your system may be used in all your partitions if it is enabled.
6.8 Performance Management Using Management Central
Management Central is an optionally installable, free component of
Operations Navigator. Be sure to select Management Central when you
install Operations Navigator. To do so, select one of the following methods to
install Operations Navigator:
• Full Install — Installs Management Central
• Custom Install — Installs Management Central, but you have to select all
subcomponents of Operations Navigator
If you already installed Operations Navigator using Typical Install or you did
not select all components during Custom Install, you can always install
Management Central later. Go to your Client Access Folder and open
Selective Setup. Use the Selective Setup Wizard to install Management
Central.
Once you install Management Central, it appears in the tree hierarchy in your
Operations Navigator window. You need to expand Management Central to
access it. For more information on how to use Management Central, refer to
the Operations Navigator home page at: http://www.as400.ibm.com/oper_nav
6.8.1 Setting Up Your Central System
Management Central allows you to manage multiple AS/400 systems from a
single AS/400 system in a TCP/IP network environment. This is the best way
to manage a single AS/400 system with logical partitions. The logical
partitions must be part of the TCP/IP network to be managed by Management
Central. The primary partition should act as the central system to manage the
logical partitions. The logical partitions are called endpoint systems in this
kind of network.
To set up the primary partition as your central system, you must have
authority to use the program that starts the Management Central server
110
Slicing the AS/400 with Logical Partitioning: A How to Guide
(program QYPSSTRS in library QSYS). The first time you start Management
Central, you must choose your central system.
To change your central system, right-click on Management Central and
select Change Central System .
6.8.1.1 Adding Logical Partitions
A logical partition acts as an endpoint system. Endpoint systems are
managed by the central system. When you first start Management Central,
right-click AS/400 Endpoint Systems and select Discover Systems. This
adds any connected AS/400 systems to your network as endpoint systems.
If you want to add logical partitions, manually as endpoint systems to your
Management Central Network, right-click AS/400 Endpoint systems.Select
New Endpoint System . Enter the system name of the specific logical
partition.
Figure 50. Adding an Endpoint System in Management Central
6.8.1.2 Creating System Groups
You can define a system group which is a collection of endpoint systems.
Endpoint systems can belong to several system groups at once. The
Capacity Planning and Performance
111
advantage of the system group is that it can be managed from the central
system just like a single system. If you add your logical partitions on your
AS/400 system to a system group, this group can be easily managed by your
primary partition.
To create a system group, complete these steps:
1. In the Operations Navigator window, open Management Central.
2. Right-click on AS/400 System Groups and select New System Group.
3. On the New System Group dialog, specify a name for the new system
group. The name of the system group must be unique.
4. Select the logical partitions from the endpoint systems listed in the
Available systems list. Click the Add button to add the logical partitions to
the Selected systems list (see Figure 51).
5. Click OK to create the new system group with your logical partitions.
Figure 51. Creating New System Group
112
Slicing the AS/400 with Logical Partitioning: A How to Guide
6.8.2 Collecting Performance Data
To collect performance data, use Management Central Collection Services.
Collected data can be used by the Performance Tools for AS/400 licensed
program product (5769-PT1). If you prefer realtime viewing of your
performance data, Management Central provides an easy-to-use graphical
interface for monitoring systems performance.
You can start Collection Services on a single system or you can start
Collection Services on system groups. Collection Services collects
information about the operation of your system that can be used to
understand response time and throughput. This performance data can be
used to make adjustments to programs and operations. It can also help you
improve response times and throughput.
Performance data can be collected by using the OS/400 Start Performance
Monitor (STRPFRMON) command. Using this command, 30 database files
are created to collect performance data. Collection Services stores your data
for each collection in a single collection object from which you can create as
many different sets of database files as you need. You can use the database
files with the Performance Tools for AS/400 licensed program (5769-PT1).
Collection Services also lets you:
• Manage your collection objects with the Operations Navigator graphical
interface.
• Collect performance data continuously.
• Control what performance data is collected and how it is used.
• Move performance data between releases without converting the data.
6.8.2.1 Starting Collection Services on a Single Partition
To start collecting system performance data, you have to start Collection
Services. Follow these steps to start Collection Services on a single partition:
1. Start Management Central in the AS/400 Operations Navigator window.
2. Expand Endpoint Systems.
3. Expand the partition where you want to collect the performance data.
4. Expand Configuration and Services.
5. Right-click Collection Services and select Start Collecting (Figure 52 on
page 114).
Capacity Planning and Performance
113
Figure 52. Start Collecting Performance Data on an Endpoint System
6. The default retention period for collection objects is one day. You may
want to specify Permanent on the General Page, if you do not want
Collection Services to delete your objects for you.
7. Click the Data to Collect tab.
8. Collection Services provides profiles to use for you. Use the Standard plus
profile to collect all the data categories typically needed by the tools in
Performance Tools for AS/400 (5769-PT1). To collect a smaller set of
categories, select Standard or Minimum .
9. Click OK to start Collection Services immediately or click Schedule to
start Collection Services at a later time.
All the performance data collected by Collection Services is stored in a single
collection object. You manage your collection objects in the Operations
Navigator graphical interface.
6.8.2.2 Starting Collection Services on a System Group
If you want to collect performance data on more than one partition, you can
start Collection Services on a system group. The system group will contain all
the logical partitions from which you want to collect performance data.
114
Slicing the AS/400 with Logical Partitioning: A How to Guide
Follow these steps to start Collection Service on a system group containing
logical partitions:
1. Start Management Central in the AS/400 Operations Navigator window.
2. Expand AS/400 System Groups.
3. Expand the partition where you want to collect the performance data.
4. Expand Configuration and Services.
5. Right-click Collection Services and select Start Collecting.
6. The default retention period for collection objects is one day. You may
want to specify Permanent on the General Page, if you do not want
Collection Services to delete your objects for you.
7. Click the Data to Collect tab.
8. Collection Services provides profiles for you to use. Use the Standard plus
protocol to collect all of the data categories typically needed by the tools in
Performance Tools for AS/400 (5769-PT1). To collect a smaller set of
categories, select Standard or Minimum .
9. Click OK to start Collection Services immediately, or click Schedule to
start Collection Services at a later time.
6.8.2.3 Customizing Data Collections
You can control what kind of data is collected by Collection Services and how
often it is collected. You can select from the standard profiles provided by
Collection Services. The profiles correspond to the settings in the OS/400
Start Performance Monitor (STRPFRMON) command for all data or for
system data.
You can select Custom to create your own customized profile from a list of
available data categories, such as System CPU, Local Response Time, Disk
Storage, and IOPs. You can specify how often data is collected for each
category of data.
For each category of data that you collect, you can specify how often the data
will be collected. For many categories, you want to select the default
collection interval, which you can set from predefined settings between 15
seconds and 60 minutes. The recommended setting is 15 minutes (see
Figure 53 on page 116).
Capacity Planning and Performance
115
Figure 53. Start Collection Services on a System Group
6.8.2.4 Creating Database Files
Database files can be created automatically using Collection Services to
collect performance data. You can also create database files from the
collection object, where the data is stored after it is collected.
If you want to create your database files automatically as data is collected,
select Create database files on the Start Collection Services dialog.
To export performance data from a collection object to database files, perform
the following steps:
1. In AS/400 Operations Navigator, expand an AS/400 system. You can
select an AS/400 endpoint system under Management Central or an
AS/400 system to which you have a direct connection under My AS/400
Connections (or your active environment). Select the logical partition
where you want the database file to be created.
2. Expand Configuration and Service.
3. Click Collection Services.
116
Slicing the AS/400 with Logical Partitioning: A How to Guide
4. Right-click the collection object you want to export to database files, and
select Create Database Files.
5. On the Create Database Files dialog (Figure 54), select the categories
from the collection object to include in the database files. You can also
select a different time period and sampling interval, as long as the
collection object contains data to support your selections.
6. Click OK .
Figure 54. Creating Database Files from Collection Services
The job to create the database files is submitted to that logical partition where
the performance data was collected. If you collected performance data for a
System Group, there will be a collection object for each system or partition in
this group. You have to create database files for each system or for each
partition separately. There are no additional tools within Management Central
for analyzing these database files.
Capacity Planning and Performance
117
6.9 Monitoring Realtime System Performance
Management Central provides a realtime performance monitor for AS/400
systems. This tool can be easily used to present realtime performance data
on logical partitions. To collect performance data for later analysis, use
Collection Services. The realtime performance monitor shows the
performance data as it happens. You cannot save this information for later
analysis.
6.9.1 Management Central
Management Central monitor graphs present system performance data in an
easy-to-use graphical interface that you can directly manipulate to get
different or more detailed data. Monitors allow you to collect performance
data simultaneously for a wide variety of system metrics, for any system or
system group, and for any length of time. Once you start a monitor, you are
free to do other tasks on your AS/400 system, in Operations Navigator, or on
your PC. In fact, you can even turn your PC off! It will continue to monitor your
systems and perform any threshold commands or actions you specified. Your
monitor will run until you decide to stop it. To effectively monitor realtime
AS/400 system performance, create a Management Central monitor.
The metrics shown in Table 15 can be selected for your monitor graphs.
Table 15. Metrics for Realtime Performance Monitor Graphs
118
Metric Name
Metric Description
CPU Utilization
This metric shows the percentage of available processing
unit time consumed by jobs on your system.
Interactive Response Time
This metric shows the response time that interactive jobs
experience on your system.
Transaction Rate
The average number of transactions per second
completed by jobs on your system.
Batch Logical Database
I/O
This metric shows the average number of logical
database input/output (I/O) operations currently
performed by batch jobs on the system.
Disk Arm Utilization
This metric shows the percentage of disk arm capacity
currently used on your system during the time you collect
the data.
Disk Storage
This metric shows the percentage of disk arm storage that
is full on your system during the time you collect the data.
Slicing the AS/400 with Logical Partitioning: A How to Guide
Metric Name
Metric Description
Disk IOP Utilization
This metric shows how busy the disk input/output
processors (IOPs) are on your system during the time you
collect the data.
Communications IOP
Utilization
This metric shows how busy the communications
input/output processors (IOPs) are on your system during
the time you collect the data.
Communications Line
Utilization
This metric shows the amount of data that was actually
sent and received on all your system communication
lines.
LAN Utilization
This metric shows the amount of data that was actually
sent and received on all your local area network (LAN)
communication lines.
Machine Pool Faults
This metric shows the number of faults per second
occurring in the system's machine pool.
User Pool Faults
This metric shows the number of faults per second
occurring in all of the user pools on the system.
6.9.1.1 Creating a New Monitor
To create a new monitor, complete these steps:
1. Expand Management Central, right-click on Monitors and select New
Monitors.
2. Select General page to:
a. Enter a unique name for your monitor. You can enter any name for your
monitor, up to 256 characters long.
b. Enter a description for your monitor. Giving your monitor a description
may help you distinguish any particular monitor from the other monitors
you have running. This can be especially useful when you have many
monitors running at once or when you want to copy monitors from one
logical partition to another.
c. Choose your metrics. You can select any metric, a group of metrics, or
all the metrics from the list to be included in your monitor. Your options
are shown in Table 15 on page 118.
Capacity Planning and Performance
119
Figure 55. Adding a New Monitor
You can choose to run a monitor that contains any combination of these
metrics on any AS/400 system, on any logical partition, or on system groups.
After you name your monitor and select metrics, you are ready to view and
change detailed metric information for each metric you selected for your
monitor.
When you create a new monitor, you can specify actions that you want to
occur on your PC when a threshold is triggered or reset. A threshold is a
setting for a metric that is collected by a monitor. Threshold actions occur on
your PC to notify you when threshold events occur. For example, you can
choose to open monitor graphs automatically when thresholds are triggered.
Threshold actions are different from any threshold commands that you may
have set. Threshold commands run on your host system or client systems,
while threshold actions occur on your PC.
When you are done working with the metric information on the New
Monitor-Metrics page, be sure to click the Edit Thresholds button. This
allows you to enable your thresholds and set threshold commands to run on
the host or client when thresholds are triggered or reset. You can also use the
120
Slicing the AS/400 with Logical Partitioning: A How to Guide
New Monitor-Actions page to set threshold actions to occur automatically on
your PC when thresholds are triggered or reset.
To edit metric information, simply select the metric that you wish to edit from
the drop-down list on the New Monitor-Metrics page (see Figure 56).
Figure 56. Edit Metric
Change any of the following properties:
• Collection interval
This information specifies how often metric information is collected.
Increase or decrease this number as needed.
• Retention period
This information specifies how long the metric information will be stored
on the host system. At the end of this period, the metric information will be
deleted from the database even if your monitor is still running.
Capacity Planning and Performance
121
• Maximum graphing value
This information specifies the highest value that will appear on the vertical
axis of the monitor graph for this metric.
• Display time
This information specifies the number of minutes that will appear on the
horizontal axis of the monitor graph for this metric.
To edit the tresholds for a selected metric, click the Edit Threshold button.
Use the threshold window to enable your thresholds. Before you can set any
threshold commands, you must turn on your thresholds by selecting the
Enable Threshold option. You can then use this window to enter any
commands you want to occur when threshold trigger and reset values are
reached.
6.9.1.2 Threshold Trigger Options
Threshold actions allow you to determine what you want to happen on your
PC when your metric reaches a certain value (called the trigger value). You
can also specify what you want to happen on your PC when your metric
reaches a second value (called the reset value). You can choose to add
events to the Event Log when your metrics reach trigger or reset values. You
can set alarms on your PC. You can even set monitor graphs to open
automatically on your PC when your metrics reach trigger or reset threshold
values. Figure 57 on page 123 shows an example of the New
Monitors-Actions tab.
122
Slicing the AS/400 with Logical Partitioning: A How to Guide
Figure 57. Treshold Trigger Options
Table 16 on page 124 provides an explanation for each action.
Capacity Planning and Performance
123
Table 16. Threshold Trigger Options
Action
Result
Open monitor
Displays the monitor graph when system performance reaches a
threshold trigger for a particular metric. This allows you to see a
graphical view of your system performance data as it is being
collected. You do not have to keep the monitor graph open on your
PC all the time. It opens automatically if you select this action, and
you can keep the monitor graph open even if you close Operations
Navigator.
Log event
Adds an entry to the Event Log on the central system indicating that
the threshold was triggered.
Open
Event Log
Displays the Event Log on your PC when a threshold trigger occurs.
Much like the Open Monitor function, this action opens the Event
Log only when you really need it.
Sound alarm
Sounds an alarm on your PC when system performance reaches a
trigger value.
6.9.1.3 Threshold Trigger and Threshold Reset
The Edit Thresholds window (Figure 58) provides a place for you to specify
the conditions that must be met to trigger and reset this threshold and the
commands that should be run when the threshold is triggered and when it is
reset. You cannot specify or change these conditions and commands unless
the Enable threshold box is checked. You cannot specify the threshold reset
conditions and commands until you have specified Value and Duration for the
Threshold trigger.
124
Slicing the AS/400 with Logical Partitioning: A How to Guide
Figure 58. Edit Thresholds
You can specify the following conditions and commands for Threshold trigger
and for Threshold reset:
• Value
Specifies the condition that must be met to trigger or to reset this
threshold.
• Duration
Specifies the number of consecutive collection intervals that the value
must meet the criteria to cause a threshold trigger or reset event.
Specifying a higher number of collection intervals for Duration helps to
avoid unnecessary threshold activity due to frequent spiking of values.
• Host command
Specifies the command to be run on the AS/400 host system when the
threshold is triggered or reset.
Capacity Planning and Performance
125
• Client command
Specifies the command to be run on the PC when the threshold is
triggered or reset. This command is run only if you are viewing this monitor
on a PC when the threshold is triggered or reset. The command is run on
every PC where you are running this monitor.
6.9.1.4 Threshold Commands
Use threshold settings to automate any command that you want to run when
thresholds are triggered or reset. For example, you can set an AS/400
command that stops any new job from starting when CPU utilization reaches
90%. You can then set another command that allows new jobs to start when
CPU utilization falls back to 70%. In another situation, you may have a
monitor that collects data on average CPU utilization for a particular system.
You can set thresholds and specify commands to keep the average CPU
utilization between 20% and 90% or any boundaries you choose.
6.9.1.5 Parameters on Host or Client Command
You can use the following parameters with host commands or client
commands to be run when a threshold is triggered or reset:
Parameter
$DATE
$MON
$RDUR
$RVAL
$INTVL
$SEQ
$TIME
$TDUR
$TVAL
$VAL
Passed Data
Date
Monitor name
Reset duration
Reset value
Collection interval
Sequence number
Time
Trigger duration
Trigger value
Current value
Consider the following example. The following host command uses the
$TIME and $TVAL parameters to pass to the program the time that the
threshold was triggered and the trigger value (see Figure 59 on page 127):
CALL LIB01/PROG02 ($TIME $TVAL)
126
Slicing the AS/400 with Logical Partitioning: A How to Guide
Figure 59. New Command Definition for Thresholds
6.9.1.6 Creating a Command Definition
Management Central enables you to define a command and then run it on
multiple systems. To do this, you must first create your command definition,
using a few quick steps:
1. In AS/400 Operations Navigator, expand Management Central.
2. Expand Definitions.
3. Right-click on Command and select New Definition.
4. Specify a name for the definition, a brief description, and the command to
be run. You can click Previous Commands to select from a list of
commands you previously ran, or you can click Check Syntax to validate
the syntax of the command on the central system.
5. To specify options concerning the job log or inquiry messages, click the
Options tab.
6. Click the Sharing tab to specify whether you want to share this package
definition with other users.
7. Click OK .
After you create your command definition, you can run the command on
systems or groups.
Capacity Planning and Performance
127
6.9.2 Running a Monitor
Once you create a monitor by selecting metrics and setting threshold actions,
the next step is to run the monitor on your AS/400 systems. To start or stop a
monitor in Management Central, select the monitor that you wish to start and
use the start and stop icons in your toolbar. You can also follow the simple
steps described here.
To start a monitor, follow these steps:
1. In Management Central, select Monitors.
2. Right-click the monitor you want to start, and select Start .
3. Select the systems or system groups on which you want to run the
monitor, and click OK .
To stop a monitor, complete these steps:
1. In Management Central, select Monitors.
2. Right-click on the monitor that you want, and select Stop.
Figure 60. Performance Monitor — CPU Utilization and Disk Storage
128
Slicing the AS/400 with Logical Partitioning: A How to Guide
When your monitor is up and running, work with the monitor graphs to see
realtime system performance data in an easy-to-use graphical interface
(Figure 60 on page 128).
You can use monitor graphs to find more information about system
performance. For example, move your mouse over a collection point (or data
point) within a graph to get information about the exact time and the AS/400
system for which the data was collected. When you click a data collection
point in a monitor graph, highly detailed information for that data point
appears in the upper right Details panel of the monitor window. The lower
right Properties panel shows properties for the information that you select in
the Details panel.
6.9.2.1 Advanced Tips for Monitor Graphs
Once you create a new monitor, you can pick up some tips to make
monitoring your system or your partitions even easier. Management Central
provides a great deal of flexibility and functionality that you can use to get the
most out of your system performance data. For example, you can:
•
•
•
•
•
Create shortcuts to your monitors
Use the Event Log
Create different graph line colors for each logical partition
Change threshold values directly on your graph
Sort the bars on your monitor graphs
The example in Figure 60 on page 128 shows you the Performance Monitor
from one system. You can easily expand this view to show more than one
system. This gives you the flexibility to watch and manage the performance
on your system with logical partitions. For example, if you want to control the
average CPU Utilization of your logical partitions, you simply have to create a
new monitor with this metric and during start you select all your partitions or,
even better, your system group with all the logical partitions. If you want to be
notified if the CPU Utilization on a logical partition is more than a specific
percentage, set a threshold, and the monitor of that partition becomes red if
the threshold is reached. In Figure 61 on page 130, the Performance Monitor
shows two logical partitions. Each partition is represented by its own line
chart. Move the mouse over the line to get detailed information about the
metrics.
Capacity Planning and Performance
129
Figure 61. Monitoring Two Logical Partitions
6.10 On-Going Performance Monitoring
After the initial installation of a multi-partitioned system, close monitoring of
the performance of each partition becomes a normal task, as with
independent physical systems. However, a new dimension is introduced with
logical partitioning.
The logically partitioned machine allows the movement of resources between
partitions. During your on-going analysis, you may find that the workload in a
three-processor partition has decreased to the point where two processors
can do the work. You may now decide to reduce the resources in that partition
and allocate them to another partition where the workload has increased.
Once the changes are implemented, you need to analyze again the
performance of both the giving and the recipient partitions to make sure that
you achieve your objective. Logical partitioning does not alter this standard
cycle to any great extent.
Another situation is where the workload on some partitions have increased,
but there is no scope for re-allocating any existing resources. In this case,
you would need to upgrade your physical machine and acquire additional
130
Slicing the AS/400 with Logical Partitioning: A How to Guide
resources such as processors, main storage, and interactive performance, to
cope with the workload. Incorporating additional resources in the existing
configuration would be done using the standard technique of resource
allocation to individual logical partition described elsewhere.
As we saw in 6.6, “Available Tools” on page 98, there is a rich array of tools
available for systems management. Each site needs to develop its own
performance management strategy. On a non-partitioned system, this
process is applied to the entire physical machine. With a partitioned machine,
it is repeated as many times as there are partitions on the physical machine.
Capacity Planning and Performance
131
132
Slicing the AS/400 with Logical Partitioning: A How to Guide
Chapter 7. Inter-Partition Communication
This chapter explains how to implement virtual OptiConnect on primary and
on secondary partitions. It explains the difference between implementing
OptiConnect/400 with external buses and implementing OptiConnect/400 on
virtual hardware. It also describes the capabilities of OptiConnect/400.
Inter-partition communication is not limited to virtual OptiConnect. You can
use existing communications adapters in each logical partition to
communicate to another logical partition. If you have Token-Ring or Ethernet
adapters in your logical partitions, you can also use all the applications
described in this chapter. However, to have the full advantage of high-speed
communications between partitions, you should use virtual OptiConnect.
7.1 OptiConnect/400 Capabilities
The fastest way to communicate between logical partitions is to use virtual
OptiConnect. To understand the methology of OptiConnect, there will be a
brief introduction to using OptiConnect/400. The only difference between
virtual OptiConnect and OptiConnect/400 is that virtual OptiConnect uses no
additional hardware. Virtual OptiConnect is communication within one
physical system divided into logical partitions. Virtual OptiConnect requires a
single license for OptiConnect/400 that can be used by all partitions.
7.1.1 OptiConnect/400
OptiConnect/400 (see Figure 62 on page 134) is a combination of hardware
and software that allows users to connect multiple high-end AS/400 systems
using a high-speed fiber optic bus connection. OptiConnect hardware
consists of OptiConnect Receiver cards that are installed in a dedicated I/O
Expansion Tower. They are connected to Satellite Systems using fiber-optic
cables.
OptiConnect software consists of the following items:
• Additions to OS/400 that provide fast path Distributed Data Management
(DDM) access across an optical bus. OptiConnect/400 is a priced feature
and must be installed as 5769SS1 (product option 23).
• APIs, if the OptiMover PRPQ (5799-FWQ) has been purchased, that can
be coded to by third parties to use the hardware.
• A connection manager that manages OptiConnect resources.
• An agent job that runs on the server on behalf of client requests.
© Copyright IBM Corp. 1999
133
Two features differentiate OptiConnect from traditional
communications-based distributed operations:
• The first is a system bus connection between multiple systems using
high-speed fiber optic technology.
• The second is an I/O driver embedded in the operating system that
streamlines the application access to data on a remote system. To
accomplish this, OptiConnect provides a shared bus on which systems
communicate using a peer-to-peer protocol. After OptiConnect establishes
system connections on the shared bus, much of the advanced
peer-to-peer communication (APPC) protocol stack is bypassed.
Hub System
Fiber Optic Links
Satellite System
Satellite System
Figure 62. OptiConnect/400 Concept
7.1.1.1 OptiConnect/400 Bandwidth
Traditional communication protocol overhead is too impractical for heavy
workload distributed applications. To minimize latency, or the time it takes to
send a message and receive a response, the protocol chosen must be
efficient in the number of code steps required. To maximize bandwidth, or the
amount of data that can be transferred in a unit of time, high-performance
hardware must be used.
The OptiConnect channel is very efficient and the best solution to both
latency and bandwidth. The length of the cable affects how low latency can
get. As distance increases, the speed of light becomes a limiting factor for a
given bandwidth. OptiConnect is limited to 500 meters over the 1063 Mbps
(megabit per second) link used on RISC AS/400 models. It is limited to two
kilometers over the 220 Mbps link supported on CISC AS/400 models.
134
Slicing the AS/400 with Logical Partitioning: A How to Guide
7.1.2 OptiMover for OS/400
OptiMover for OS/400 is a special version of OptiConnect for OS/400. It
provides application programs with Application Program Interfaces (APIs) to
the services of the OptiConnect Connection Manager and Device Driver.
With the OptiMover for OS/400 PRPQ, applications may use the high-speed
optical bus to communicate among systems on the shared bus, including the
object transfer functions in ObjectConnect, but the high-speed DDM
capabilities of OptiConnect are not available. Because OptiMover is
essentially the same as the OS/400 OptiConnect feature, it uses the same
combination of hardware and software that allows you to connect multiple
AS/400 systems on a shared optical bus.
Note: The PRPQ number for OptiMover is P84291, program 5799-FWQ.
7.2 OS/400 Functions Using OptiConnect/400
The AS/400 has many functions available in its operating system that can use
OptiConnect/400. The most common are Distributed Data Management/400
(DDM) and ObjectConnect/400 for concurrent save and restore of objects
without using a tape media.
7.2.1 Distributed Data Management (DDM)
DDM is part of the OS/400. The DDM support of the AS/400 system allows
application programs or users to access data files that reside on remote
systems, and starting with OS/400 V4R4, also on logical partitions.
Distributed Data Management provides horizontal growth using
OptiConnect/400 as a simple method of accessing files on other systems
through the use of DDM files. There is a large overhead on CPU and
response times using traditional DDM with common communication methods.
IBM offers DDM at system bus speed by using OptiConnect/400.
Inter-Partition Communication
135
Figure 63. Distributed Data Management Concept
Figure 63 shows that there is an overhead using Distributed Data
Management over a communication line. OptiConnect/400 uses a fast Device
Driver (DD) and no ICF/APPC function is involved in a DDM transaction.
There are some limitations when implementing DDM of which you should be
aware. Customers should read and understand these limitations which can be
found in Distributed Data Management, SC41-3307.
Some of the most common commands are shown in the following list along
with the details on how DDM affects them:
• ALCOBJ
When using the Allocate Object (ALCOBJ) command over DDM, the
object on the source system is locked as well as on the target side. When
DDM-related files are being allocated, a longer time is required to
complete the command.
• CHKOBJ
Using the Check Object (CHKOBJ) CL command to check the existence of
members across DDM does not work.
• Multimembered files
DDM files do not have the option to access multiple members. You can do
this by specifying the member name in the Override with Database File
(OVRDBF) command before accessing the file.
136
Slicing the AS/400 with Logical Partitioning: A How to Guide
• Query/400
You can run an existing query across DDM. However, no changes can be
made to it. The product Query/400 should not be used. If you need to run
queries against DDM files, perform a Passthrough or Telnet to the other
system. Then, run the query with all its options from there.
• Remote data areas and remote data queues
Remote data areas and remote data queues are supported by
OptiConnect/400. They can be used, but there must be an existing
communications path from one partition to the other.
• SBMRMTCMD
You can use the Submit Remote (SBMRMTCMD) command, but no
attributes are passed over DDM.
• SQL
SQL splits its workload between the two partitions. There are no problems
using SQL over DDM. If you plan to use SQL across DDM, perform a
benchmark first to predict the performance.
7.2.2 ObjectConnect/400
ObjectConnect/400 provides six specialized save and restore commands.
With these commands, you get the ability to save and restore objects,
libraries, DLOs, and IFS directories to a second partition. ObjectConnect/400
operates over several communication methods. Virtual OptiConnect always
provides the best performance out of all of them. Some of the most common
commands are described in the following list:
• SAVRST
You can use the Save/Restore (SAVRST) command to save one or more
objects in directories, send them to another system, and restore them. It
can also save entire directories (not to be confused with entire systems).
The SAVRST command supports the same options as the SAV command.
• SAVRSTOBJ
You can use the Save/Restore Object (SAVRSTOBJ) command to save
one or more objects, send them to another system, and restore them. The
SAVRSTOBJ command supports the same options as the SAVOBJ
command, including the use of the OMITOBJ parameter.
• SAVRSTCHG
You can use the Save/Restore Changed Objects (SAVRSTCHG)
command to save one or more changed objects, send them to another
Inter-Partition Communication
137
system, and restore them. An example of this would be a situation where
you want to maintain duplicate sets of files on two different systems. The
SAVRSTCHG command supports most of the same options as the
SAVCHGOBJ command, including the use of OMITOBJ parameters. You
may use the OMITLIB parameter with this command. You may also specify
generic values for the LIB parameter on this command.
• SAVRSTLIB
You can use the Save/Restore Library (SAVRSTLIB) command to save
one or more libraries, send them to another system, and restore them. The
SAVRSTLIB command supports the same options as the SAVLIB
command, including the use of the OMITLIB and OMITOBJ parameters.
You may also use generic values for the *LIB parameter on this command.
• SAVRSTDLO
You can use the Save/Restore Document Library Object (SAVRSTDLO)
command to save one or more document library objects, send them to
another system, and restore them. The SAVRSTDLO command supports
the same options as the SAVDLO command.
• SAVRSTCFG
You can use the Save/Restore Configuration (SAVRSTCFG) command to
save one or more configuration objects, send them to another system, and
restore them. The SAVRSTCFG command supports most of the options
and parameters as the SAVCFG and RSTCFG commands do.
When you copy your configuration by using the SAVRSTCFG command,
the system saves and restores the following object types:
– *CFGL Configuration List
– *COSD Class-of-Service Description
– *CNNL Connection List
– *MODD Mode Description
– *NTBD NetBIOS Description
– *IPXD IPX Description
7.2.3 Remote Journal Function
The remote journal function allows you to establish journals and journal
receivers on a remote AS/400 system that are associated with specific
journals and journal receivers on a local system. The remote journal function
can replicate journal entries from the local system to the journals and journal
receivers that are located on the remote system after they are established.
138
Slicing the AS/400 with Logical Partitioning: A How to Guide
The remote journal function helps to efficiently replicate journal entries to one
or more systems. You can use the remote journal function with application
programs to maintain a replica database. A replica database is a copy of an
original database that resides on another AS/400 system. An original
database resides on a primary system. Applications make changes to the
original database during normal operations. Previous to V4R2M0, you could
accomplish a similar function by using the Receive Journal Entry (RCVJRNE)
command. The remote journal function replicates journal entries to the
remote system at the Licensed Internal Code layer. Moving the replication to
this lower layer provides the following advantages:
• The remote system handles more of the replication overhead
• The overall system performance and journal entry replication performance
are improved.
• An option is available to have replication occur synchronously to the
operation that is causing the journal entry deposit.
• Journal receiver save operations can be moved to the remote system.
In 7.4.2, “Remote Journal Function” on page 159, you can find a detailed
description of the remote journal function.
7.3 Virtual OptiConnect/400
Virtual OptiConnect/400 (see Figure 64 on page 140) uses the same software
as OptiConnect/400, but there is no need for any additional hardware units
and I/O cards on a system with logical partitions. Virtual OptiConnect
emulates the real OptiConnect hardware between logical partitions. In
addition, beginning with OS/400 V4R4, you can also run the TCP/IP protocol
over both types of OptiConnect.
There are no hardware requirements for virtual OptiConnect. Virtual
OptiConnect emulates the real hardware.
If you use virtual OptiConnect, you have the same software requirements as
with OptiConnect. You must purchase and install one of the following options:
• 5769SS1 product option 23
• OptiMover PRPQ (5799-FWQ)
Note: You only need a single license of OptiConnect across all partitions. You
do not need to purchase OptiConnect with all partitions.
Inter-Partition Communication
139
Figure 64. Virtual OptiConnect/400 Concept
If you enabled inter-partition communication using virtual OptiConnect during
partition setup, you can communicate with each logical partition.
Beginning with OS/400 V4R4, IBM has significantly reduced the license fee
for OptiConnect/400. We recommend that you re-consider the applicability of
the OptiConnect/400 solution in your environment again if price was the only
issue that prevented you from using this unique technology.
7.3.1 Transferring Data between Logical Partitions
Virtual OptiConnect uses Direct Memory Access functions to transfer pages
from one partition to another. Figure 65 on page 141 shows an example of a
data transfer from partition to partition. The data resides on the DASD in
partition 0 and has to be written to the DASD on partition 1. The following
process occurs:
1. A read request arrives at the DASD IOP on partition 0. Data is read and
transferred through IOP using a Direct Memory Access (DMA) function
into memory pages in partition 0.
2. Pages are forwarded, using DMA, into an intermediate buffer on partition
0.
3. To get the data to the memory of partition 1, once again, a DMA is used.
Pages are transferred from the intermediate buffer on partition 0 into
memory on partition 1. No communication link is used for this memory
copy. Virtual OptiConnect emulating the real OptiConnect hardware gets
140
Slicing the AS/400 with Logical Partitioning: A How to Guide
the data from the buffer on partition 0 to the memory of partition 1 by using
a virtual OptiConnect IOP.
4. The pages are transferred from the memory of partition 1 through the IOP
for the DASD on that partition and written to the disk. This is done again
by a DMA function.
Using this kind of mechanism, virtual OptiConnect provides a real high-speed
connection between logical partitions. The only likely bottleneck for data
transfer is the IOP and DASD arms in the partitions.
Figure 65. Read Request Using Virtual OptiConnect
7.3.2 Setting Up Virtual OptiConnect over SNA
OptiConnect can be used by any application that was written to use
Distributed Data Management (DDM). Many applications that use an AS/400
database can transparently use DDM without changes to the application.
Applications that access a database by DDM can use different ways to do so:
• Traditional APPC DDM (SNA)
• OptiConnect/400 (SNA and TCP/IP)
• Virtual OptiConnect (SNA and TCP/IP)
There are two different ways to route data requests through OptiConnect. If
you use a special keyword in the DDM file, you will use the Fastpath
OptiConnect. Using this method, OptiConnect agent jobs start in the
Inter-Partition Communication
141
OptiConnect Connection Manager and run in the QSOC subsystem. Although
Fastpath OptiConnect is the faster means of communications, you should be
aware that two-phase commit is not supported by this method.
The second method is called Extended Function Path OptiConnect. Using
this method, the OptiConnect agent jobs are started by the advanced
program-to-program communications (APPC) attach manager and run in the
QCMN subsystem. Two-phase commit is supported by Extended Function
Path OptiConnect.
7.3.2.1 Setting Up Fastpath OptiConnect
If you install the OptiConnect/400 software, the QYCTSOC device description
is created automatically. This device always remains varied off. This device
description is necessary for OptiConnect, and therefore, must not be deleted.
Whenever an APPC conversation is directed to this device, the OptiConnect
device driver redirects the conversation through the OptiConnect bus.
To create a DDM file for Fastpath OptiConnect, perform the following steps:
1. Enter the CRTDDMF command and press F4.
2. Enter the DDM file name and library. This is the name that the application
must know to have access to the data.
3. Enter the remote file name and library. In this library of the remote system,
the data file is actually stored. In an environment with logical partitions.
this is the name and library for the data file in that logical partition.
4. For the Remote Location parameter, specify the system name of the target
system , where the database file is located. To find out the name, use the
Display Network Attributes ( DSPNETA) command on the remote system. In
an environment with logical partitions, this is not the name of the logical
partition. It is the name of the system as specified in the logical partition.
5. Press F10 for additional parameters.
6. On the Additional Parameters display, page down and enter QYCTSOC for the
APPC device description. Whenever DDM finds a DDM file with this
special keyword used in the file description, OptiConnect routes requests
using the Fastpath method.
7. In the Mode parameter, you can specify any valid mode of your AS/400
system. If you want to run the OptiConnect Agent jobs to start with
USRPRF specified in the QYYCDTSU job description, then you must use
the same keyword as on the APPC device description, QYCTSOC.
142
Slicing the AS/400 with Logical Partitioning: A How to Guide
Figure 66. Creating a DDM File to Use OptiConnect (Part 1 of 2)
Figure 67. Creating a DDM File to Use OptiConnect (Part 2 of 2)
Inter-Partition Communication
143
7.3.2.2 Setting Up Extended Function Path Routing
For such applications as SNA distribution services and the remote journal
function, you need to set up Extended Function Path Routing. This enables
you to route data requests through OptiConnect without the special device
keyword, QYCTSOC. You must create OptiConnect controllers and devices of
type *OPC on the logical partitions.
Figure 68. Matching Parameters for Extended Function Path Routing
First, you must create a pair of *OPC controllers, one on each of the logical
partitions that uses OptiConnect to communicate to another logical partition.
One controller on a logical partition must have *PRI (primary) defined in the
Data Link Role. The other controller must be *SEC (secondary). The LAN
DSAP parameter value must be valid and identical for both controllers on
both partitions in the pair. Valid values are 04, 08, 0C, 10, 14, .... 78, and 7C.
Use the following command to configure the *OPC controller on a logical
partition:
CRTCTLAPPC CTLD(name) LINKTYPE(*OPC) RMTSYSNAME(sysname) ROLE(*PRI or *SEC)
DSAP(##)
Figure 69 and Figure 70 on page 145 show an example of creating two *OPC
controllers on partition LPAR1 and on partition LPAR2.
144
Slicing the AS/400 with Logical Partitioning: A How to Guide
Figure 69. Create Controller Description on LPAR1 to Connect to LPAR2
Figure 70. Create Controller Description on LPAR2 to Connect to LPAR1
Inter-Partition Communication
145
Now, create device descriptions for each controller on each partition. Specify
APPN(*NO), because the *OPC controller only accepts devices that are non
APPN-capable.
Use the following command to configure the *OPC controller on a logical
partition:
CRTDEVAPPC DEVD(name) RMTLOCNAME(LPAR2OPC) ONLINE(*NO)
LCLLOCNAME(LPAR1OPC) APPN(*NO)
Figure 71 and Figure 72 on page 147 show an example of creating two device
descriptions to attach to the *OPC controller.
Figure 71. Create Device Description on LPAR1 to Attach LPAR2CTL
146
Slicing the AS/400 with Logical Partitioning: A How to Guide
Figure 72. Create Device Description on LPAR2 to Attach LPAR1CTL
These devices do not actually describe a device. They are used to control
authority and access to the OptiConnect path.
To enable requests over OptiConnect using the created controllers and
devices, you must start the QSOC subsystem first. After starting that
subsystem, you can vary on all *OPC controllers and devices. The
OptiConnect path is completely established when both controllers and
devices change to the ACTIVE state.
These *OPC controllers can be used to provide APPC communication
capability over virtual OptiConnect. Some of the applications for that interface
are:
• An application program using the inter communication file (ICF) file
interface
• The communication programming interface (CPI) communication call
interface
• The customer information control system (CICS) file interface
Inter-Partition Communication
147
To create a DDM file for Extended Function Path Routing, complete these
steps:
1. Enter the CRTDDMF command, and press F4.
2. Enter the DDM file name and library. This is the name that the application
must know to have access to the data.
3. Enter the remote file name and library. In this library of the remote system,
the data file is actually stored. In an environment with logical partitions,
this is the name and library for the data file in that logical partition.
4. For the Remote Location parameter, specify the system name of the target
system , where the database file is located. To find out the name, use the
Display Network Attributes ( DSPNETA) command on the remote system. In
an environment with logical partitions, this is not the name of the logical
partition. It is the name of the system as specified in the logical partition.
5. Press F10 for additional parameters.
6. On the Additional Parameters display, page down. Enter QYCTSOC for the
APPC device description. Whenever DDM finds a DDM file with this
special keyword used in the file description, OptiConnect routes requests
using the Fastpath method.
In the Mode parameter, you can specify any valid mode of your AS/400
system. If you want to run the OptiConnect Agent jobs starting with USRPRF
specified in the QYYCDTSU job description, then you must use the same
keyword as on the APPC device description, QYCTSOC.
7.3.3 Starting Virtual OptiConnect
To start virtual OptiConnect, you have to start subsystem QSOC. The best
way to do this, is by adding a command to your startup program that will be
run at IPL time.
To start the QSOC subsystem, you have to enter the Start Subsystem
(STRSBS) command on each partition:
STRSBS QSOC/QSOC
When you start the QSOC subsystem, the OptiConnect connection manager,
SOCMGR, starts as an autostart job.
7.3.4 Ending Virtual OptiConnect
You can end OptiConnect by ending the QSOC subsystem. Before you end
OptiConnect on a logical partiton, make sure that there are no application
programs still using this connection.
148
Slicing the AS/400 with Logical Partitioning: A How to Guide
If you use the remote journal function on this partition, end this function
before ending the QSOC subsystem. remote journal function jobs are not
displayed with the Work with Active Jobs (WRKACTJOB) command.
To end virtual OptiConnect, enter the following command:
ENDSBS QSOC *IMMED
If you ended the QSOC subsystem on one logical partition, you still can use
virtual OptiConnect on your system if you have more partitions that have
enabled Interpartition OptiConnect.
If you configured Extended Function Path Routing by creating controllers and
devices, vary off the *OPC controllers and the corresponding controllers on
the other logical partition. These controllers are in an unusable state when
you end the QSOC subsystem. You have to vary off and then vary on those
controllers to activate them.
7.3.5 Determining Virtual OptiConnect Activity
Use the following commands to verify the OptiConnect activity and to obtain
information about its resources and components.
• Work with Active Jobs (WRKACTJOB)
• Work with OptiConnect Activity (WRKOPCACT)
• Display OptiConnect Link Status (DSPOPCLNK)
7.3.5.1 Work with Active Jobs (WRKACTJOB)
You can monitor OptiConnect activity by using the Work with Active Jobs
(WRKACTJOB) command. To see the active jobs running in the QSOC
subsystem, enter the following command:
WRKACTJOB SBS(QSOC)
You will see a SOCMGR job if the QSOC subsystem is running. On the target
partition, you will also see one or more agent jobs (SOCAnnnnnn jobs)
running in susbsytem QSOC.
Agent jobs are active or inactive (prestarted). To determine if an agent job is
active, enter a 5 in front of the agent job in the WRKACTJOB display. Then
choose option 11 (Display call stack) or option 14 (Display open files).
Inactive agent jobs have no files open.
Inter-Partition Communication
149
Active agents jobs are present until one of the following actions occur:
• The source partition job ends or the user logs off.
• The source partition job ends or the user uses the Reclaim Resources
(RCLRSC) command.
• The source partition job ends or the user uses the Reclaim DDM
Conversations (RCLDDMCNV) command.
Figure 73 shows a sample of the Work with Active Jobs display.
Figure 73. Work with Active Jobs — QSOC Subsystem
7.3.5.2 Work with OptiConnect Activity (WRKOPCACT)
This is the most useful command in an OptiConnect environment. Using the
Work with OptiConnect Activity (WRKOPCACT) command, you get
information about the bus activity. You can also change the state of
OptiConnect resources. Plus, you can see information about database
transactions using the virtual OptiConnect bus and the connection status of
all involved client and server partitions.
150
Slicing the AS/400 with Logical Partitioning: A How to Guide
The Work with OptiConnect Activity provides three different views:
• The first view you see when you enter the WRKOPCACT command. It
shows the OptiConnect connectivity from the perspective of an application
partition. Here is the only place where you can see the number of
transactions between logical partitions. You see the connection status of
virtual OptiConnect to any other logical partition. The collection period is
shown on the top of the display.
• The next view shows this partition as a client. You can access this view by
pressing F11 (Client Statistics View).
• The next view shows this partition as a server. You can access this view by
pressing F11 (Server Statistics View).
Figure 74 shows the Work with OptiConnect Activity when you enter the
following command:
WRKOPCACT
Figure 74. Work with OptiConnect Activity
The Active/degraded connection status indicates that this system has an
operational connection to the system named in the System column through
the adapter resource. This connection is in use by user or agent jobs.
Inter-Partition Communication
151
The degraded status means redundancy is reduced because this resource
has one or more unusable local or remote links. To determine the reason for
the resource degraded status, check the status of both the local and remote
links. A link status other than Ready or Active indicates that the link is
unusable and will result in the degraded status. A blank connection status
indicates that the link does not exist and will not show the degraded status.
7.3.5.3 Display OptiConnect Link Status (DSPOPCLNK)
The Display OptiConnect Link Status (DSPOPCLNK) command shows the
connection status between systems in the fiber-optic network. In an
environment with logical partitions, you can use this command to check the
status of the connection. Most of the other views of this command do not
apply to virtual OptiConnect. There are no real OptiConnect hardware
resources in a virtual OptiConnect Environment.
The Top Link, for example, shows the fiber optic connection that runs from the
top port of the System Unit Optical Link Processor to the top port of the
OptiConnect Adapter. The Bottom Link shows the fiber optic connection that
runs from the bottom port of the System Unit Optical Link Processor to the
bottom port of the OptiConnect Adapter. The Redundant Link shows the fiber
optic connection that runs from the top port of one OptiConnect Adapter to
the bottom port of the other OptiConnect Adapter. In an environment with
virtual OptiConnect, the top link and the bottom link are shown as Unknown
and the Redundant link is left blank.
7.3.6 Setting Up OptiConnect/400 over TCP/IP
Prior to OS/400 Version 4 Release 4, you could only use the SNA protocol to
communicate over OptiConnect. Beginning with V4R4, you can also have
applications that use Transport Control Protocol/Internet Protocol (TCP/IP) to
communicate over OptiConnect. The AS/400 provides a standard IP
interface, which allows existing applications and services to work unchanged
by defining a TCP/IP interface that uses OptiConnect.
The IP over OptiConnect implementation allows applications that use a
TCP/IP Socket interface to communicate over the OptiConnect shared bus
when running in an AS/400 cluster. This allows the growing set of
applications that are distributed across multiple AS/400 systems (three-tiered
applications) to take advantage of the high bandwidth and low latency of the
OptiConnect shared bus. Prior to this implementation, the OptiConnect
shared bus was available using a proprietary API (OptiMover) or using an
APPC controller. The TCP/IP socket interface to OptiConnect provides the
highest bandwidth, with the lowest latency communication method between
AS/400 systems and between logical partitions. The IP over OptiConnect
152
Slicing the AS/400 with Logical Partitioning: A How to Guide
implementation provides a line driver interface between the bottom of the
TCP/IP stack and the OptiConnect device driver. It provides communication
services over a shared bus in an AS/400 cluster.
You can configure up to four IP interfaces to be used by OptiConnect. Each of
the interfaces must be on a separate subnet. To configure this interface, you
can use the standard AS/400 methods, which are Configure TCP/IP
(CFGTCP) and the Operations Navigator graphical user interface (GUI).
Once you configure the IP interface, you have to start it. It allows you to
communicate to other partitions that have enabled Inter-partition
OptiConnect. To do so, those partitions must have an interface with the same
subnet. All protocols which use Internet Protocol (IP), including Transmission
Control Protocol (TCP), User Datagram Protocol (UDP), and so forth.
7.3.6.1 Defining the OptiConnect Interface
To add a new interface to the Transmission Control Protocol/Internet Protocol
(TCP/IP), you can use the Add TCP/IP Interface (ADDTCPIFC) command. If you
add a new interface, which will be used by virtual OptiConnect, you do not
have any hardware resources that you are really using. Therefore, the
interfaces created by the ADDTCPIFC command for virtual OptiConnect are
logical interfaces.
7.3.6.2 Configuring the OptiConnect Interface
You can configure the TCP/IP interfaces for OptiConnect in two different
ways. In the first configuration, the OptiConnect bus has to be seen similar to
a LAN.
Figure 75. Configuration of the OptiConnect Interface — LAN Version
In Figure 75, you see an example for an environment with logical partitions.
We have three logical partitions and all are Inter-partition OptiConnect
enabled. To configure the interface for use by OptiConnect, use the Add
Inter-Partition Communication
153
TCP/IP Interface (ADDTCPIFC) command. Each *OPC interface has to be
assigned a unique IP address within the subnet. Even if it seems that there
are two interfaces for LPAR2, it is only one interface with the IP address
9.1.1.16. This interface is attached to the optical bus (OptiConnect). You will
see a new parameter special value here, *OPC. This special value is used if
you are adding an OptiConnect interface over TCP/IP.
The required parameters are:
INTNETADR (Internet address)
Specifies an Internet address that the local system responds to on
this interface. An interface is associated with a line description.
The Internet address is specified in the form nnn.nnn.nnn.nnn,
where nnn is a decimal number ranging from 0 through 255. An
Internet address is not valid if it has a value of all binary ones or all
binary zeros for the network identifier (ID) portion or the host ID
portion of the address. If the Internet address is entered from a
command line, the address must be enclosed in apostrophes.
LIND (Line description)
Specifies the name of the line description associated with the new
interface. The line description must be defined before the TCP/IP
interface can be added. In an OptiConnect environment, you must
use the special value *OPC here.
SUBNETMASK (Subnet mask)
Specifies the subnet mask, which is a bit mask that defines the
part of the network where this interface attaches. The mask is a
32-bit combination that is logically ANDed with the Internet
address to determine a particular subnetwork. The bits of the
mask set to the value one (1) determine the network and
subnetwork portions of the address. The bits set to the value zero
(0) determine the host portion of the address.
LCLIFC (Associated local interface) — Not a required parameter
Use this parameter to associate the interface that you are
currently defining with an existing local TCP/IP interface. The
associated local interface (LCLIFC) is used to allow transparent
subnetting (also known as Proxy Arp) between the associated
interfaces, to define Frame Relay unnumbered networks or for
load balancing.
To configure the TCP/IP interfaces for this example, enter the following
command statement on partition LPAR1:
ADDTCPIFC INTNETADR(’9.1.1.15’) LIND(*OPC) SUBNETMASK(’255.555.255.0’)
154
Slicing the AS/400 with Logical Partitioning: A How to Guide
Figure 76. ADDTCPIFC — Add TCP/IP Interface for OptiConnect
To configure the interfaces on LPAR2 and LPAR3, enter the following
command statement for partition LPAR2:
ADDTCPIFC INTNETADR(’9.1.1.16’) LIND(*OPC) SUBNETMASK(’255.555.255.0’)
For partition LPAR3, enter this command statement:
ADDTCPIFC INTNETADR(’9.1.1.17’) LIND(*OPC) SUBNETMASK(’255.555.255.0’)
Another approach for configuring TCP/IP is to use the associated local
interface parameter. If you use this method, there is no reason to define new
subnets for the OptiConnect bus. No external route tables need to be updated
to provide connectivity between the OptiConnect interfaces and the rest of
the TCP/IP network. You use the OptiConnect interfaces as part of existing
local subnets to which the AS/400 system is attached.
Each OptiConnect interface defines an endpoint of a point-to-point
OptiConnect connection between two AS/400 systems. Configuring the
interface configured this way is actually not an interface in the system or
partition to which you are currently signed on. Instead, it is an existing
interface in the endpoint of that connection. The existing local interface for
the TCP/IP network will be specified as the associated local interface to the
OptiConnect interface.
Inter-Partition Communication
155
In Figure 77, you see an environment with three logical partitions. All of these
partitions have their own interface to the Token-Ring network.
Figure 77. Configuring OptiConnect Interface (Part 1 of 2)
Logical partition LPAR1 has the Internet address 9.1.1.11, LPAR2 has the
Internet address 9.1.1.12, and LPAR3 has the Internet address 9.1.1.13. The
subnet mask for these three partitions is 255.255.255.192.
Now we want to add the OptiConnect interfaces for this example. You can see
the new configuration in Figure 78.
Figure 78. Configuring OptiConnect Interface (Part 2 of 2)
156
Slicing the AS/400 with Logical Partitioning: A How to Guide
On partition LPAR1, perform the following steps:
1. Define the endpoint of the OptiConnect connection for logical partition
LPAR2 using the associated local interface: 9.1.1.11
The Internet address for the endpoint in logical partion LPAR2 is: 9.1.1.12
2. Define the endpoint of the OptiConnect connection for logical partition
LPAR3 using the associated local interface: 9.1.1.11
The Internet address for the endpoint in logical partition LPAR3 is:
9.1.1.13
Use the following commands for the configuration in LPAR1:
1. For the endpoint in partition LPAR2, enter the following command
statement:
ADDTCPIFC INTNETADR(’9.1.1.12’) LIND(*OPC)
SUBNETMASK(’255.555.255.255’) LCLIFC(’9.1.1.11’)
2. For the endpoint in partition LPAR3, enter the following command
statement:
ADDTCPIFC INTNETADR(’9.1.1.13’) LIND(*OPC)
SUBNETMASK(’255.555.255.255’) LCLIFC(’9.1.1.11’)
No local interfaces for the OptiConnect bus have to be configured. The
subnet mask in this configuration method must always be 255.255.255.255 for
the interfaces describing the endpoint of an OptiConnect connection.
You need to perform these configurations on partition LPAR2 and on partition
LPAR3. From the view of LPAR2, you have to configure the endpoints in
partitions LPAR1 and LPAR3. From the view of partition LPAR3, you have to
configure the endpoints in partitions LPAR1 and LPAR2.
Use the following commands for the configuration in LPAR2:
1. For the endpoint in partition LPAR1, enter:
ADDTCPIFC INTNETADR(’9.1.1.11’) LIND(*OPC)
SUBNETMASK(’255.555.255.255’) LCLIFC(’9.1.1.12’)
2. For the endpoint in partition LPAR3, enter:
ADDTCPIFC INTNETADR(’9.1.1.13’) LIND(*OPC)
SUBNETMASK(’255.555.255.255’) LCLIFC(’9.1.1.12’)
Inter-Partition Communication
157
Use the following commands for the configuration in LPAR3:
1. For the endpoint in partition LPAR1, enter:
ADDTCPIFC INTNETADR(’9.1.1.11’) LIND(*OPC)
SUBNETMASK(’255.555.255.255’) LCLIFC(’9.1.1.13’)
2. For the endpoint in partition LPAR2, enter:
ADDTCPIFC INTNETADR(’9.1.1.12’) LIND(*OPC)
SUBNETMASK(’255.555.255.255’) LCLIFC(’9.1.1.13’)
Another advantage of this method is that, if one of the OptiConnect paths
becomes inactive, packets are automatically routed over the backup
interface.
If subsystem QSOC has been started and the TCP/IP interfaces are active,
all the traffic between the logical partitions are routed over the OptiConnect
path.
7.4 Application Examples Using Virtual OptiConnect/400
If you configured your AS/400 system for using logical partitions, you have
the capability to use virtual OptiConnect without needing new hardware. You
need OS/400 product option 23 (OptiConnect), and you need to enable the
Inter-partition OptiConnect during configuration of the logical partitions.
These are the hardware and software requirements for the application
examples described in the following sections.
7.4.1 SNA Distribution Services (SNADS)
If you want to use the OptiConnect communications path between AS/400
logical partitions for SNA Distribution Services, you have to configure *APPC
controllers. You can find information on how to configure *APPC controllers
and devices in 7.3.2.2, “Setting Up Extended Function Path Routing” on page
144.
You need to create a directory entry, a routing table entry, and a distribution
queue. The definition of the distribution queue has to use the names of the
local location and the remote location as defined in the *APPC device
description.
7.4.1.1 Adding a Directory Entry
You need to add an *ANY entry for the logical partition in your directory. To
add this entry, enter this command statement:
ADDDIRE USRID(xxx/*ANY) USRD(yyy) SYSNAME(zzz)
158
Slicing the AS/400 with Logical Partitioning: A How to Guide
In this statement, note the following explanation:
• xxx is the name of the remote system.
• yyy is the user description.
• zzz is the name of the remote system.
7.4.1.2 Adding a Distribution Queue
You need to add a distribution queue for the logical partition. To add the
queue, enter the following command statement:
ADDDSTQ DSTQ(xxx) RMTLOCNAME(yyy)
Note the following explanation:
• xxx is the name of the queue.
• yyy as specified in the *APPC device.
Then, enter the following statement:
DSTQTYPE(*SNADS)
MODE(*NETATR)
Or, specify a mode.
RMTNETID(*NETATR)
LCLLOCNAME(zzz)
In this statement, zzz as specified in the *APPC device.
7.4.1.3 Creating a Routing Table
You need to add a routing table for the logical partition. Enter this statement:
ADDDSTRTE SYSNAME(xxx)
Here, xxx is the name of the remote system. Note these options:
FAST(yyy)
STATUS(zzz)
DATAHIGH(aaa)
DATALOW(bbb)
7.4.2 Remote Journal Function
The remote journal function was introduced in OS/400 with V4R2. This
function offers a fast and reliable method to transfer journal receiver data to a
remote AS/400 system. With the availability of logical partitions on the
AS/400 system, the remote journal function is ideal for data replication in a
logical partition environment.
You can use the remote journal function with application programs to maintain
a replica of the local database. A replica database is a copy of the original
Inter-Partition Communication
159
database that resides on another system or on another partition. Before the
introduction of the remote journal function, you had to use the Receive
Journal Entry (RCVJRNE) command to receive journal entries. A RCVJRNE
exit program then receives these entries from the journal receiver and sends
them to the remote system or logical partition. This could be done by any
available communications method.
For more information on this function, refer to the redbook AS/400 Remote
Journal Function for High Availability and Data Replication, SG24-5189.
7.4.3 Data Replication Using Journals
There are a lot of high availability solutions in the marketplace using journals
for data replication. Solutions that are available from IBM Business Partners,
use local journaling and the Receive Journal Entry (RCVJRNE) command. In
this environment, database changes are generated by user's applications
running on a source production system. These database changes create
journal entries in a local journal receiver. This is a common functionality used
by applications on AS/400 systems.
These journal entries in a local journal receiver must be transferred to the
remote system. The high availability solutions provided by IBM business
partners use the Receive Journal Entry (RCVJRNE) command and an exit
program to transfer the journal entries to the target system. The
communications transport can be done using one of the available
communication lines to the target system. The fastest communication method
is virtual OptiConnect. The data is transmitted asynchronously to the target
system. On the target system these journal entries are retrieved and usually
stored in a user space. Another job, or many jobs, replay the changes into a
copy of the source database. Once the journal entries are applied to the
database, you have an exact copy of the production database on the target
system.
The disadvantage of this solution is:
• The asynchronous transfer of journal receiver entries means data latency.
If your source systems fails while there are some journal entries still
waiting for transmission, a few final database transactions are trapped on
the source system. The database replica on the target system lags behind
the production database on the source system.
• The data passes through many layers of the system software. It crosses
the Machine Interface (MI) boundary several times. This means extra CPU
cycles and therefore increased CPU utilization.
160
Slicing the AS/400 with Logical Partitioning: A How to Guide
In Figure 79, you can see a high-level description of this process.
Figure 79. Hot-Backup Environment without Remote Journal Function
7.4.4 Data Replication Using the Remote Journal Function
To get rid of the disadvantages, data latency and increased CPU utilization,
the remote journal function provides a much more efficient method of journal
entry transport. Efficient low-level system code is used to capture and
transmit journal entries directly from the source system to associated journals
and journal receivers on a target system. There is no need to buffer the
resulting journal entries on the production system. Much of the processing is
done below the MI. The CPU utilization decreases using this method. If the
remote journal function was activated in synchronous mode, the journal
entries are replicated to the target system's main storage before the
production system's database gets updated. The data latency is driven to
zero.
There are still more tasks in a high availability solution on the AS/400 system
than the replication of the database. In fact, the available high availabilty
solutions can take advantage of this more efficient transport mechanism. In
customer environments for a hot-backup scenario, there is a need for
providing the required management facilities. The available solutions for
hot-backup support replication for more system objects than the database.
In Figure 80 on page 162, you can see a high-level figure for a hot-backup
solution using the remote journal function.
Inter-Partition Communication
161
Figure 80. Hot-Backup with Remote Journal Function
7.4.5 Setting Up the Remote Journal Function
The remote journal function is built upon the local journal function. Before the
remote journal function can be performed, the local journal function on the
source system must first be set up.
Complete the following steps to setup the remote journal function:
1. Set up a relational database (RDB) directory entry for the remote system.
The RDB directory contains database names and values that are
translated into communications network parameters. In an RDB directory
entry, specify the communications protocol and the corresdonding
communications path. In Figure 81 on page 163, you can see an example
of how to add an RDB entry using virtual OptiConnect.
You must add one entry for each remote RDB. The RDB must be unique
within the RDB directory and within the distributed network.
162
Slicing the AS/400 with Logical Partitioning: A How to Guide
Add RDB Directory Entry (ADDRDBDIRE)
Type choices, press Enter.
Relational database . . . . . . > LPAR2
Remote location:
Name or address . . . . . . . > LPAR2OPC
Type . . . . . . . . . . . . . *SNA
*SNA, *IP
Text . . . . . . . . . . . . . . > 'Remote RDB on Partition LPAR2'
Device:
APPC device description
Local location . . . . . .
Remote network identifier
Mode . . . . . . . . . . .
Transaction program . . .
F3=Exit F4=Prompt
F24=More keys
.
.
.
.
.
.
.
.
.
.
. > LPAR2OPC
. > PRIMEOPC
. > *NONE
. *NETATR
. *DRDA
F5=Refresh
F12=Cancel
Name, *LOC
Name, *LOC, *NETATR
Name, *LOC, *NETATR, *NONE
Name, *NETATR
Character value, *DRDA
Bottom
F13=How to use this display
Figure 81. Add RDB Directory Entry — SNA
2. If you use TCP/IP over OptiConnect for the transfer of the remote
journalentries, you have to change the DDM server job attributes for
TCP/IP. The DDM server should start on IPL.
You need to specify whether client systems are required to have a
password in addition to a user ID on incoming connection requests to this
partition as a server. This parameter takes effect on the next DRDA or
DDM connect request over TCP/IP.
Change DDM TCP/IP Attributes (CHGDDMTCPA)
Type choices, press Enter.
Autostart server . . . . . . . .
Password required . . . . . . .
*YES
*YES
*NO, *YES, *SAME
*NO, *YES, *ENCRYPTED, *SAME
Figure 82. Change DDM TCP/IP Attributes
3. Create the local journal and the local journal receiver on the production
system by using the Create Journal Receiver ( CRTJRNRCV) command and
the Create Journal (CRTJRN) command.
Inter-Partition Communication
163
4. Start local journaling using the Start Journal Physical File (STRJRNPF)
command.
5. Add the remote journal using the Add Remote Journal (ADDRMTJRN)
command. This command associates the remote journal on the target
system with the specified journal on the source system.
6. Before journal entries can be replicated to a remote journal, the remote
journal must be activated. To activate a remote journal, use the Change
Remote Journal ( CHGRMTJRN) command. This command must be run on the
source system, on which the source journal is located. Each remote
journal must be activated individually by using this command.
Using the Change Remote Journal (CHGRMTJRN) command, you decide
which mode you use for the replication of the journal entries.
7.4.6 Benefits of the Remote Journal Function
The remote journal function is well suited on a logical partition environment.
You can offload workload from the production partition to the backup partition.
The remote journal function replicates journal entries to the remote partition
at the Licensed Internal Code layer. Moving the replication to this lower layer
provides these advantages:
• The remote partition handles more of the replication overhead.
• The overall system performance and journal entry replication performance
is improved.
• An option is available to have replication occur synchronously to the
operation that causes the journal deposit. Using this mode, journal entries
are replicated to the main memory on the remote partition first. After an
arrival confirmation is returned to the source partition, the journal entry is
deposited to the local receiver. Next, the actual database update is made.
• Journal receiver save operations can be moved to the remote system.
7.4.7 Using Your Database 24 Hours a Day, 7 Days a Week
If you need to provide 24-hour by 7-day operation of your AS/400 system, you
have to consider a solution that gives you the ability to run your business as
well as take care of the housekeeping of your vital data. Figure 84 on page
166 shows an example of how the remote journal function can help in such an
envirnonment.
164
Slicing the AS/400 with Logical Partitioning: A How to Guide
Figure 83. Using Remote Journal Function for 24-by-7 Database Availability
The journal entries are transferred to the backup partition using the remote
journal function. The changes are then applied to the backup copy of the
application so it reflects what is on the production partition. It is then possible
to save from the backup partitions copy of the application, which would
ensure that the production partition is available to the end users.
7.4.8 DB2 Multisystem
DB2 Multisystem is a parallel processing technique that provides an almost
unlimited scalability option for databases. Using DB2 Multisystem, you have
the capability to attach multiple AS/400 systems (up to 32 systems) together
in a "shared nothing" cluster. Shared nothing means that each system in the
coupled network owns and manages its own main memory and disk storage.
Once the AS/400 systems are connected, database files can be spread
across the storage units on each connected system. To users, the file looks
like a local file on their system. From the user's perspective, the database
appears as a single database. The user can run queries in parallel across all
the systems in the network and have realtime access to the data in the files.
To install DB2 Multisystem, use option 27 in the list of installable options for
the OS/400 operating system.
Inter-Partition Communication
165
Figure 84. Distribution of Database Files across Logical Partitions
In Figure 84, the database is spread across three logical partitions. Running
a query against a database file allows each partition that contains a part of
the database to run its portion of the query at the same time as the other
partitions. Once each partition has run its part of the query over its local data,
the results are combined and returned to the user. Each partition processes a
smaller amount of data, which results in the overall query completing in a
much shorter amount of time.
166
Slicing the AS/400 with Logical Partitioning: A How to Guide
Appendix A. Logical Partitioning Terminology
This appendix reviews most of the common hardware terms used within the
logical partitioning context.
A.1 Processor
All machines have at least one processor that drives the primary partition.
Through logical partitioning, multi-processor (N-way) machines may be
configured with secondary partitions. Each partition contains one or more
processors.
A.2 Main Storage
Main storage is comparable to RAM on a personal computer. Sufficient main
storage must be allocated to primary and secondary partitions to ensure
optimal performance.
A.3 Bus
A bus is a piece of hardware used by the AS/400 system to bind its individual
physical parts into a single machine. It is like a highway used by the
processor to gain access to all its parts. Within the logical partitioning
function, buses are some of the components allocated to logical partitions to
create independent systems within the physical machine.
Resources can be allocated at a bus level. This means that all resources
attached to the bus are allocated to a single bus at a time.
A.4 Input/Output Processor (IOP)
The input/output processor (IOP) is a piece of hardware that allows devices,
such as disks and tape drives, or communications lines, to attach to a bus,
using an input/output adapter (IOA). They connect to the system like this: the
device is connected to the IOA, which is connected to the IOP, which is
connected to the bus.
Resources can be allocated at an IOP level. In this case, the bus to which the
IOP is attached is shared, and all the resources attached to the IOP are
allocated to a single partition at a time. It is then possible to switch this IOP
from partition to partition as the need arises.
© Copyright IBM Corp. 1999
167
A.5 Multifunction IOP (MFIOP)
This is a type of input/output processor (IOP) that can connect to a variety of
different input/output adapters (IOA). For example, an MFIOP can support
disk units, a console, and a tape drive.
There is a second type of multifunction IOP called a combined multifunction
IOP (CFIOP). Conceptually, they are the same. However, there are some
special rules on how they attach to a bus and are allocated to logical
partitions.
A.6 Load Source
The load source is a special disk that contains the base software for the
machine, called System Licensed Internal Code (SLIC). It also contains
information describing how the logical partitions are configured.
Logical partitions represent independent systems. As such, each partition
must have a Load Source disk containing its own System Licensed Internal
Code and logical partitioning configuration data. One of the actions required
in the creation of logical partitions is to define the disk unit which will become
the Load Source.
A.7 System Licensed Internal Code (SLIC)
The System Licensed Internal Code or SLIC is the microcode that drives the
hardware. On a partitioned machine, each partition has its own SLIC loaded
on to the Load Source. This allows partitions to run on different levels of
system software if necessary.
A.8 Partition Licensed Internal Code (PLIC)
The Partition Licensed Internal Code (PLIC) is a thin layer of microcode,
which is common to all logical partitions in a partitioned machine. It contains
the code required to implement logical partitioning across the physical
machine.
A.9 Hypervisor
The Hypervisor is made of a small portion of the primary partition SLIC and
the PLIC, which runs across all partitions. It is the enabling agent for logical
partitioning.
168
Slicing the AS/400 with Logical Partitioning: A How to Guide
A.10 Removable Media Device
Every logical partition must have either a tape drive or an optical CD-ROM
device. This is used to load software (for example, the System Licensed
Internal Code) and data into the partition. Such devices are known as
removable media devices.
A.11 Console
Each logical partition must have a console attached to it through an IOP.
Through this console, each logical partition looks and functions like an
independent system. The console allows you to enter commands and view
the operation of a logical partition.
A primary partition console allows you to see and manage many aspects of
all secondary partitions. A secondary partition console allows you access
only to its attached partition.
A.12 Expansion Unit
Expansion units are racks attached to a system to accommodate features
and devices external to the system unit. Some expansion units can support
disk only. Others can support a variety of IOPs, including disk and tape
controllers.
The IOPs on an expansion unit access the system unit using a bus that
connects the expansion unit to the system unit.
Expansion units may be needed in the creation of logical partitions to provide
support for all of the resources required.
A.13 Ownership and Sharing
Logical partitioning allows the switching of some resources between
partitions. This can be done at the bus level or the IOP level.
A bus can be owned in two ways: dedicated or shared. In a dedicated mode,
the bus belongs exclusively to one partition. In a shared mode, the IOPs
attached to the bus can be switched from partition to partition.
Logical Partitioning Terminology
169
The permutations used for these terms are:
• Own bus dedicated (indicating a non-shareable resource)
• Own bus shared (indicating a shareable resource)
• Use bus shared (indicating that a partition is sharing a bus owned by
another partition)
Accordingly, there are two choices available when partitioning a machine:
• Bus level partitioning, where a bus along with all the IOPs attached to it is
dedicated to a single partition. In this case, there is no sharing of IOP
resources.
• IOP level partitioning, where a bus is shared among partitions and the
IOPs attached to it can be independently switched from partition to
partition as the need arises.
A.14 Addition and Removal
Addition and removal are the terms used to switch resources between
partitions. When resources are removed from a logical partition, they become
available for addition to another logical partition.
Some addition and removal actions may require an entire system restart.
Other such actions may only require a secondary logical partition restart.
A.15 OptiConnect
OptiConnect is a function that allows you to connect multiple systems using a
physical high-speed, fiber-optic bus. Virtual OptiConnect emulates the
physical OptiConnect hardware to achieve the same objective between
logical partitions.
To use Virtual Opticonnect, all you need is to purchase the software license
for OptiConnect or OptiMover and configure your logical partitions to enable
the Virtual OptiConnect function.
In addition to Virtual OptiConnect, inter-partition communication can also be
achieved with standard LAN/WAN facilities. In such a case, each partition
needs to have a dedicated communication adapter.
170
Slicing the AS/400 with Logical Partitioning: A How to Guide
Appendix B. Sample RPG and C Programs
The following examples show how to use the unblocked MI instruction,
MATMATRI, to retrieve partition resource information.
B.1 Sample ILE RPG Program — Output to Spooled File
0001.00
0002.00
0003.00
0004.00
0005.00
0006.00
0007.00
0008.00
0009.00
0010.00
0011.00
0012.00
0013.00
0014.00
0015.00
0016.00
0017.00
0018.00
0019.00
0020.00
0021.00
0022.00
0023.00
0024.00
0025.00
0026.00
0027.00
0028.00
0029.00
0030.00
0031.00
0032.00
0033.00
0034.00
0035.00
0036.00
0037.00
0038.00
0039.00
0040.00
0041.00
0042.00
0043.00
0044.00
0045.00
0046.00
0047.00
0048.00
0049.00
0050.00
0051.00
0052.00
© Copyright IBM Corp. 1999
H*********************************************************************
H* COMPILE THIS PROGRAM WITH DFTACTGRP(*NO)
*
H*
*
H* Sample ILE RPG program using the MI instruction MATMATR to
*
H* materialize partition information on the AS/400.
*
H*********************************************************************
F*
F* Print file specifications for output
F*
FQSYSPRT
O
F 132
PRINTER
F*
D*
D* Prototype of MI MATMATR instruction
D*
DMatMAtr
PR
EXTPROC('_MATMATR1')
D
*
VALUE
D
2
CONST
D*
D* Working variables for Materialize
D*
DAttribPtr
s
*
inz(%addr(Attributes))
DMatOption
s
2
inz(x'01E0')
D*
D* Receiver variable for Materialize
D*
DAttributes
DS
512
D BytPrv
10i 0
D BytAvl
10i 0
D NumParts
3u 0
D CurPart
3u 0
D PriPart
3u 0
D Reserved
5
D LogSerial
10
D Reserved2
6
D MinProc
5u 0
D MaxProc
5u 0
D CurProc
5u 0
D Reserved3
2
D CfgMinMem
10u 0
D CfgMaxMem
10u 0
D CurAvlMem
10u 0
D MinIntPerf
3u 0
D MaxIntPerf
3u 0
D CurIntPerf
3u 0
D Reserved4
1
D*
D Description
S
25
inz(' ')
D ValueN
S
10 0 inz(0)
D ValueC
S
10
inz(' ')
C*
C* Set Bytes Provided to size of Recevier Variable (Attributes
C*
171
0053.00
0054.00
0055.00
0056.00
0057.00
0058.00
0059.00
0060.00
0061.00
0062.00
0063.00
0064.00
0065.00
0066.00
0067.00
0068.00
0069.00
0070.00
0071.00
0072.00
0073.00
0074.00
0075.00
0076.00
0077.00
0078.00
0079.00
0080.00
0081.00
0082.00
0083.00
0084.00
0085.00
0086.00
0087.00
0088.00
0089.00
0090.00
0091.00
0092.00
0093.00
0094.00
0095.00
0096.00
0097.00
0098.00
0099.00
0100.00
0101.00
0102.00
0103.00
0104.00
0105.00
0106.00
0107.00
0108.00
0109.00
0110.00
0111.00
0112.00
0113.00
0114.00
0115.00
172
C
C*
C*
C*
C
C*
C*
C*
C
C*
C*
C*
C
C*
C*
C*
C
C
C
C*
C*
C*
C
C
C
C*
C*
C*
C
C
C
C*
C*
C*
C
C
C
C*
C*
C*
C
C
C
C*
C*
C*
C
C
C
C*
C*
C*
C
C
C
C*
C*
C*
C
C
C
C*
C*
eval
BytPrv = %size(Attributes)
Use MATMATR MI instruction
callp
MatMAtr(AttribPtr: MatOption)
Determine if information returned
if
BytAvl >= 56
Read though the detail and output it to the printer
Except
Header
Total number of partitions that have been created
eval
eval
Except
Description = 'Total Partitions Created'
ValueN = NumParts
DetailN
Current partition
eval
eval
Except
Description = 'Current Partition Number'
ValueN = CurPart
DetailN
Logical Serial Number
eval
eval
Except
Description = 'Logical Serial Number
ValueC = LogSerial
DetailC
'
Description = 'Minimum No Processors
ValueN = MinProc
DetailN
'
Description = 'Maximum No Processors
ValueN = MaxProc
DetailN
'
Description = 'Current No Processors
ValueN = CurProc
DetailN
'
Description = 'Minimum Amount Memory
ValueN = CfgMinMem
DetailN
'
Description = 'Maximum Amount Memory
ValueN = CfgMaxMem
DetailN
'
Minimum No Processors
eval
eval
Except
Maximum No Processors
eval
eval
Except
Current No Processor
eval
eval
Except
Minimum Amount Memory
eval
eval
Except
Maximum Amount Memory
eval
eval
Except
Current Amount Memory
Slicing the AS/400 with Logical Partitioning: A How to Guide
0116.00
0117.00
0118.00
0119.00
0120.00
0121.00
0122.00
0123.00
0124.00
0125.00
0126.00
0127.00
0128.00
0129.00
0130.00
0131.00
0132.00
0133.00
0134.00
0135.00
0136.00
0137.00
0138.00
0139.00
0140.00
0141.00
0142.00
0143.00
0144.00
0145.00
0146.00
0147.00
0148.00
0149.00
0150.00
0151.00
0152.00
0153.00
0154.00
0155.00
0156.00
0157.00
0158.00
0159.00
0160.00
0161.00
C*
C
eval
Description = 'Current Amount Memory
'
C
eval
ValueN = CurAvlMem
C
Except
DetailN
C*
C* Minimum Interactive Performance %
C*
C
eval
Description = 'Minimum Interactive Perf'
C
eval
ValueN = MinIntPerf
C
Except
DetailN
C*
C* Maximum Interactive Performance %
C*
C
eval
Description = 'Maximum Interactive Perf'
C
eval
ValueN = MaxIntPerf
C
Except
DetailN
C*
C* Current Interactive Performance %
C*
C
eval
Description = 'Current Interactive Perf'
C
eval
ValueN = CurIntPerf
C
Except
DetailN
C
Except
End
C*
C
else
C
eval
ValueC = ' '
C
eval
Description = '****Error Occurred*******'
C
Except
DetailC
C
Except
End
C
endif
C*
C* exit sample program
C*
C
eval
*inlr = '1'
C
return
C*
OQSYSPRT
E 1
HEADER
2 03
O
24 'Partition Information'
O
E
DetailN
1
O
Description
28
O
ValueN
3
45
O
E
DetailC
1
O
Description
28
O
ValueC
47
O
E
END
2
O
28 '*** END OF REPORT ***'
B.2 Sample C Program — Output to Screen
#include <stdio.h>
#include <QSYSINC/MIH/MATMATR>
_MMTR_Template_T
machine_attributes;
int main(void)
{
machine_attributes.Options.Template_Size = sizeof(_MMTR_Template_T);
matmatr ( &machine_attributes, _MMTR_LPAR_INFO) ;
printf ("Total Partitions Created: %d \n",
Sample RPG and C Programs
173
machine_attributes.Options.Data.LparInfo.num_of_partition);
printf ("Current Partition Number: %d \n",
machine_attributes.Options.Data.LparInfo.partition_id);
printf ("Primary Partition number: %d \n",
machine_attributes.Options.Data.LparInfo.prim_partition);
printf ("Logical Serial Number
: %s \n",
machine_attributes.Options.Data.LparInfo.logical_ser_num);
printf ("Minimum No Processors
: %i \n",
machine_attributes.Options.Data.LparInfo.min_processors);
printf ("Maximum No Processors
: %i \n",
machine_attributes.Options.Data.LparInfo.max_processors);
printf ("Current No Processors
: %i \n",
machine_attributes.Options.Data.LparInfo.cur_processors);
printf ("Minimum Amount Memory
: %i \n",
machine_attributes.Options.Data.LparInfo.cur_min_memory);
printf ("Maximum Amount Memory
: %i \n",
machine_attributes.Options.Data.LparInfo.cur_max_memory);
printf ("Current Amount Memory
: %i \n",
machine_attributes.Options.Data.LparInfo.cur_avail_memory);
printf ("Minimum Interactive Perf: %i \n",
machine_attributes.Options.Data.LparInfo.min_percentage_int_work);
printf ("Maximum Interactive Perf: %i \n",
machine_attributes.Options.Data.LparInfo.max_percentage_int_work);
printf ("Current Interactive Perf: %i \n",
machine_attributes.Options.Data.LparInfo.cur_percentage_int_work);
}
174
Slicing the AS/400 with Logical Partitioning: A How to Guide
Appendix C. Accessing LPAR Functions
Table 17 explains the logical partitioning menu options available from the DST
and SST screens in both the primary and secondary partitions. The initial
logical partitioning menu can be reached by selecting option 11 from the main
DST menu or option 5 from the main SST menu.
Table 17. Logical Partitions Menu Options
Logical Partitioning Function
Primary
Partition
Secondary
Partition
DST
SST
DST
SST
Option 1 - Display partition information
Y
Y
Y
Y
1. Display partition status
Y
Y
Y(1)
Y(1)
2. Display partition processing configuration
Y
Y
Y(1)
Y(1)
3. Display allocated resources
Y
Y
Y(1)
Y(1)
4. Display available resources
Y
Y
Y
Y
5. Display system I/O resources
Y
Y
N
N
6. Display partition release level
Y
Y
N
N
7. Display secondary partition reference code history
Y
Y
N
N
Option 2 - Work with partition status
Y
Y
Y(1)
Y(1)
Option 3 - Work with partition configuration
Y
Y(2)
N
N
Option 4 - Recover configuration data
Y
Y
Y
Y(3)
1. Recover primary partition configuration data
Y
N
N
N
2. Update configuration data
Y
Y
N
N
3. Clear non-configured disk unit configuration data
Y
N
Y(1)
N
4. Clear non-reporting logical partitioning resources
Y
N
N
N
5. Accept load source disk unit
Y
N
Y(1)
N
6. Copy configuration data to other side
Y
N
Y(1)
N
© Copyright IBM Corp. 1999
175
Logical Partitioning Function
Primary
Partition
Secondary
Partition
7.Clear configuration data
Y
N
N
N
Option 5 - Create a new partition
Y
N
N
N
Notes:
Options with an "N" in the SST column for the primary partition are displayed. If they are
selected, an error message is sent.
Options with an "N" in either the DST or SST columns of the secondary partitions are not
displayed with the exception of note 3 (below).
1. You can only work with or display information related to this secondary partition.
2. Options 1, 2, and 10 are displayed but can only be selected when in DST of the
primary partition.
3. This menu option is displayed, but an error message is sent if selected.
Security Level Required to Access the Menu Options
You must use either the QSECOFR or 22222222 ID to access the menu options
from within DST. To access the menu options in SST, the user profile that you
use must have *SERVICE specified as one of its special authorities.
176
Slicing the AS/400 with Logical Partitioning: A How to Guide
Appendix D. Operations Console Connection for Partitioned AS/400
There are four types of console devices commonly found in an AS/400
system. These devices are:
•
•
•
•
Twinaxial workstation
ASCII workstation
Client Access console on PC
Operations Console on PC
However, logical partitions on the AS/400 system only support the twinaxial
console and the Operations Console. This appendix offers tips and hints on
how to set up Operations Console in an AS/400 partition.
D.1 What Operations Console Is
The Operations Console feature comes standard with Client Access for
Windows 95, 98, and NT V3R2M0 or Client Access Express for Windows
V4R4M0. It also requires OS/400 V4R3M0 or later release on the AS/400
system.
AS/400 Operations Console allows you to use a PC with Windows 95 or
Windows NT to access and control, either remotely or locally, the AS/400
console and control panel. Each AS/400 partition must have a console
attached to it.
Local Controlling System
Operations Console cable
AS/400 Logical Partitions
- OS/400 V4R4M0
Operations Console
running on Windows 95/98/NT
- Client Access for Win 95/NT V3R2M0
or Client Access for Windows V4R4M0
Figure 85. Operations Console Attachment
The Operations Console PC is directly connected to an AS/400 system
through a special Operations Console cable (p/n 97H7557, feature #0367 for
© Copyright IBM Corp. 1999
177
#2721 attachment, p/n 97H7556, feature #0328 for #2699 attachment and
#97H7555 for #2609 or #2612 attachment). It is connected between the PC
serial port and the AS/400 communication adapter (port 0 if a two-line
communication adapter). This PC is called Local Controlling System (LCS).
The LCS can be configured to access the AS/400 console function on either
the primary partition or the secondary partition. The LCS can also be
configured to access the AS/400 control panel function on the primary
partition. In this case, the PC must also be connected to the AS/400 control
panel port through a special Panel cable (p/n 97H7584, feature #0380 for
SPD bus system unit and p/n 97H7591, feature #0381 for PCI bus system
unit) and a second serial port on the PC. There is no control panel support on
the LCS if it is attached to the secondary partition. Only one LCS can be
connected to one AS/400 partition.
The LCS does not use the previous SNA-over-Async type of communication,
like the Client Access console, between the AS/400 system and the LCS.
Instead, the LCS acts as a router between the AS/400 system and the
Personal Communication 5250 emulation using TCP/IP through the TCP/IP
loop back function on the PC. The 5250 emulation for the AS/400 console
function can be run on another PC through a dial-up connection to the LCS.
This remote PC is called Remote Controlling System (RSC) . To use RSC for
remote access, the LCS must be a Windows NT (workstation) PC.
D.2 Direct-Attached Operations Console in Secondary Partition
To attach the Operations Console to an AS/400 partition, either the primary
partition or a secondary partition, you need to perform the following steps.
You can refer to Operations Console Setup, SC41-5508, for the details.
178
Slicing the AS/400 with Logical Partitioning: A How to Guide
Tips
Please check the following items before the Operations Console is started:
1. Ensure that the Operations Console cable is attached correctly.
2. Install the AS/400 Operations Console Connection null modem in your
Windows 95, 98, or NT PC.
3. Select the console IOP as the default ECS resource in partition
configuration.
4. Select the Operations Console as the default console type in the DST.
5. If you have twinaxial IOA attached to the same console IOP, you need
to switch off the attached twinaxial workstation with the addresses port
0 address 0, port 0 address 1 and port 1 address 0 in order to use the
Operations Console as the system console.
The following list highlights the areas that you need to pay attention to in
LPAR environment. Complete these steps:
1. Connect the Operations Console cable correctly.
Normally, port 0 is the left-hand port of feature code #2721 when it is
attached to system bus 1 of 620/S20/720. However, it will be the
right-hand port of feature code #2721 if the #2721 is attached to system
bus 2 of Model 620/S20/720. It is because the #2721 on the system bus 2
is turned over when compared to the position on the system bus 1.
Port 0 is the lower port for the communication IOA, such as #2699, for the
SPD card.
2. Install Client Access for Windows 95/NT V3R2 or the Client Access
Express for Windows V4R4 with Operations Console on your PC.
You need to configure the AS/400 Operations Console null modem (see
Figure 86 on 180).
Operations Console Connection for Partitioned AS/400
179
Figure 86. AS/400 Operations Console Modem
If you have not installed the null modem, you can install the AS/400
Operations Console null modem from X:\Program Files\IBM\Client
Access\Aoc\inf. Here, X is the directory where you installed the Client
Access base code.
3. Identify the communication IOP or the combined function IOP or the
MFIOP that has the Operations Console attached:
• #2809 (combined function IOP with #2721 attached) for PCI bus
system
• #2629 (communication IOP with #2699 attached) for SPD bus
• #2623 (communication IOP with #2609 attached) for SPD bus
• MFIOP
4. Select (or tag) the console IOP in the partition configuration Figure 87 on
181).
180
Slicing the AS/400 with Logical Partitioning: A How to Guide
Select Console Resource
System: SYSTEMA
Level of detail to display . . . . . . *ALL *ALL, *BUS, *IOP, *IOA, *DEV
Type option, press Enter.
1=Select console IOP 2=Select alternate console IOP
I/O Resource
Opt Description
_ System Bus
Serial
Part
Type-Model Number
Number
00-0000000
More...
* Indicates load source.
F3=Exit
F10=Display logical address
F9=Do not filter resource capability
F12=Cancel
Figure 87. Select Console IOP with Filtering
The communication IOP, combined function IOP, or MFIOP to which the
Operations Console is attached, is not shown as seen in Figure 87 on 181.
The IOP is filtered out by the built-in filter rules since it has no twinaxial
workstation IOA attached. You need to press F9 (Do not filter resource
capability) to get back the communication IOP, combined function IOP, or
the MFIOP for your selection as shown in Figure 88 on 182.
Operations Console Connection for Partitioned AS/400
181
Select Console Resource
System: SYSTEMA
Level of detail to display . . . . . . *ALL *ALL, *BUS, *IOP, *IOA, *DEV
Type option, press Enter.
1=Select console IOP 2=Select alternate console IOP
I/O Resource
Opt Description
_ System Bus
1
Combined Function IOP
*
Serial
Part
Type-Model Number
Number
00-0000000
2809-001
53-8303211 0000021H5312
More...
* Indicates load source.
F3=Exit
F10=Display logical address
F9=Filter resource capability
F12=Cancel
Figure 88. Select Console IOP without the Filtering
Then, select the IOP with the Operations Console PC attached.
5. Select the console IOP as the Electronic Customer Support IOP also.
To have the AS/400 partition correctly pick up the communication adapter,
you must select the console IOP in the previous step as your default
Electronic Customer Support Resource.
Go to the Work with Partition Configuration display (Figure 89 on 183),
and enter option 9 next to the secondary partition, for example, LPAR2.
182
Slicing the AS/400 with Logical Partitioning: A How to Guide
Work with Partition Configuration
System:
Type option, press Enter.
7=Select console resource 8=Select alternate IPL resource
9=Select default electronic customer support resource
10=Delete partition
SYSTEMA
Partition
Option Identifier Name
__
0
PRIMARY
9_
1
LPAR2
F3=Exit
F11=Work with partition status
F12=Cancel
F23=More options
Figure 89. Work with Partition Configuration Menu
The < sign next to the Combined Function IOP in Figure 90 indicates that
the IOP has already been selected as the console IOP.
Select Default Electronic Customer Support Resource
System: RCHASM25
Level of detail to display . . . . . . *ALL *ALL, *BUS, *IOP, *IOA, *DEV
Type option, press Enter.
1=Select IOP
I/O Resource
Opt Description
_ System Bus
1
Combined Function IOP
*<
Serial
Part
Type-Model Number
Number
00-0000000
2809-001
53-8303211 0000021H5312
More...
* Indicates load source.
F3=Exit
F10=Display logical address
F9=Do not filter resource capability
F12=Cancel
Figure 90. Select Default Electronic Customer Support Resource Menu
Select the console IOP as the default Electronic Customer Support IOP.
Operations Console Connection for Partitioned AS/400
183
6. Configure the Operations Console on your PC for the connection.
Refer to Chapter 37 of Client Access for Windows 95/NT Setup V3R2,
SC41-3512, for the details.
7. Start the Operations Console and power on the secondary partition.
8. When the Operations Console DST window pops up on the PC, use any
one of the DST user IDs to sign on:
• QSECOFR — Security capability
• 22222222 — Full capability
• 11111111 — Basic capability
Figure 91. Operations Console DST Sign On Window
Use the basic user ID 11111111 to sign on to the Operations Console.
9. If this is a first time installation, continue with the disk formatting and SLIC
installation.
10.Select Operations Console as your console device in DST.
You should now be at the IPL or Install the System menu as shown in
Figure 92 on 185.
a. Select option 3 (Use Dedicated Service Tools (DST)) and sign on as
QSECOFR.
184
Slicing the AS/400 with Logical Partitioning: A How to Guide
IPL or Install the System
System:
LPAR2
Select one of the following:
1.
2.
3.
4.
5.
Perform an IPL
Install the operating system
Use Dedicated Service Tools (DST)
Perform automatic installation of the operating system
Save Licensed Internal Code
Selection
3
Licensed Internal Code - Property of IBM 5769-999 Licensed
Internal Code (c) Copyright IBM Corp. 1980, 1999. All
rights reserved. US Government Users Restricted Rights Use duplication or disclosure restricted by GSA ADP schedule
Contract with IBM Corp.
Figure 92. IPL or Install the System Menu
b. Select option 5 (Work with DST environment). See Figure 93.
Use Dedicated Service Tools (DST)
System:
LPAR2
Select one of the following:
1.
2.
3.
4.
5.
6.
7.
8.
Perform an IPL
Install the operating system
Work with Licensed Internal Code
Work with disk units
Work with DST environment
Select DST console mode
Start a service tool
Perform automatic installation of the operating system
10. Work with remote service support
11. Work with system partitions
Selection
5
F3=Exit
F12=Cancel
Figure 93. DST Main Menu
c. Select option 2 (System devices). See Figure 94 on 186.
Operations Console Connection for Partitioned AS/400
185
Work with DST Environment
System:
LPAR2
System:
LPAR2
Select one of the following:
1.
2.
3.
4.
Active service tools
System devices
DST user profiles
System values
Selection
2
F3=Exit
F12=Cancel
Figure 94. Work with DST Environment Menu
d. Select option 6 (Console mode). See Figure 95.
Work with System Devices
Select one of the following:
1.
2.
3.
4.
5.
6.
Printers
Tape devices
Diskette devices
Optical devices
Alternate installation device
Console mode
Selection
6
F3=Exit
F12=Cancel
Figure 95. Work with System Devices Menu
e. Select option 2 (Operations Console) and press Enter.
186
Slicing the AS/400 with Logical Partitioning: A How to Guide
Select Console Type
System:
LPAR2
Select one of the following:
1. Local Console
2. Operations Console
Selection
2
F12=Cancel
Figure 96. Select Console Type Menu
11.Exit from DST and continue the system installation.
Operations Console Connection for Partitioned AS/400
187
188
Slicing the AS/400 with Logical Partitioning: A How to Guide
Appendix E. Interactive and Batch CPW Determination
If you need to consolidate a number of AS/400 systems into one AS/400
partition, you must evaluate both the interactive and batch performance
requirements for the combined workload. The methodology for the
performance determination or capacity planning is no different from the
consolidation of two AS/400 systems into one system.
This appendix provides the steps to evaluate both interactive and batch
performance requirements of the combined workload using the BEST/1 tool.
E.1 Overview of the Methodology
In general, you can characterize the system workload in many different ways.
For example, you can characterize the system workload according to the
business function that the jobs perform. You may have Order Entry, Accounts
Receivable, or Accounts Payable workloads during the day time, and
Customer Master Update and General Ledger Update workloads at night. Or,
you can simply characterize the system loading into interactive workload
during the day time and batch workload at night.
Use the following steps to determine both the interactive and the batch
performance requirements:
1. Define or characterize the workloads.
2. Define the performance expectation. It can be interactive response time,
batch runtime or the number of orders processed.
3. Collect the performance data of the workloads using the AS/400
performance monitor, STRPFRMON command.
4. Analyze the performance data and create the BEST/1 model for the
workloads. Workload can be classified according to the user ID,
subsystem name, job name, job type, or job number. You can refer to the
BEST/1 manual or Chapter 4 of the redbook AS/400 Server Capacity
Planning, SG24-2159, for details.
5. Compare the measured result with the customer’s performance
expectation.
6. Combine the workloads if they run concurrently and project the machine
requirement in terms of the CPU usage, main storage size, the expected
performance, and the partition configuration.
7. Derive the interactive CPW number and batch CPW number.
© Copyright IBM Corp. 1999
189
E.2 Interactive Performance Determination
You should have performed step 1 to step 5 in E.1, “Overview of the
Methodology” on page 189, before you can continue to determine the
interactive CPW requirement. Now you have a clear idea of your current
performance level and whether the current performance level matches the
expectation. Use the following steps to combine the interactive workloads
and evaluate them on a user defined AS/400 partition model in BEST/1:
1. Create a user-defined model based on the target AS/400 system and the
proposed LPAR configuration. You can refer to 6.7, “Creating a New
Hardware Feature in BEST/1” on page 99, for details.
2. If you want to combine two or more workloads into one AS/400 partition,
you need to save the workload in the BEST/1 model that you created.
For example, you created two BEST/1 models, INTER01 in library QGPL
and M53SINT in library QPFR_TIM53, for the interactive workloads of two
different AS/400 systems that you want to consolidate.
The following steps help you to combine the workloads into the base
model. The base model is one of the BEST/1 models that you created for
the consolidation. It is the starting base for the system performance
projection or evaluation. It is normally the biggest system or the one with
most of the communication lines and LAN adapters among the AS/400
systems. Let us assume the base model is M53SINT in library
QPFR_TIM53. You need to save the workloads from INTER01 and add
them to the model M53SINT.
a. Start AS/400 Performance Tools and work with the BEST/1 model that
you want to save the workload.
b. Choose option 5 to work with the model INTER01.
190
Slicing the AS/400 with Logical Partitioning: A How to Guide
Work with BEST/1 Models
Library . . . . .
QGPL
Name
Type options, press Enter.
1=Create 3=Copy 4=Delete
Opt Model
5
INTER01
5=Work with
6=Print
7=Rename
Text
Date
Time
Interactive Performance
02/04/99 14:51:30
Bottom
Command
===>
F3=Exit F4=Prompt
F15=Sort by model
F5=Refresh
F9=Retrieve F12=Cancel
F16=Sort by text F19=Sort by date and time
Figure 97. Choosing Option 5 to Work with the Model INTER01
c. Save the workloads of the current model.
d. Choose option 1 (Work with workloads) to work with the workloads of
the model INTER01.
Work with BEST/1 Model
Performance data . . . :
Model/Text . . . . . . :
QPFRDATA (OCTEST1)
INTER01
Interactive Performance
Select one of the following:
1. Work with workloads
2. Specify objectives and active jobs
5. Analyze current model
6. Analyze current model and give recommendations
7. Specify workload growth and analyze model
10. Configuration menu
11. Work with results
More...
Selection or command
===> 1
F20=More BEST/1 options
F21=Basic user level
F24=More keys
Figure 98. Choosing Option 1 to Work with the Workloads of the Model INTER01
Interactive and Batch CPW Determination
191
e. Save the workloads to the base model. The base model is normally the
largest system among the systems that you want to consolidate.
f. Choose option 8 to select the required workloads.
Work with Workloads
Model/Text:
INTER01
Interactive Performance
Type options, press Enter.
1=Create 2=Change 3=Copy 4=Delete
8=Save workload to workload member
Opt
8
8
8
8
Workload
Text
CLIENTAC4
INTERACTIV
NONINTER
QDEFAULT
Interactive
Interactive
Interactive
Interactive
F3=Exit F6=Add saved workload
F13=Combine workloads
5=Display 6=Print
9=Edit transactions
performance
performance
performance
performance
data
data
data
data
from
from
from
from
SG
SG
SG
SG
7=Rename
system
system
system
system
F9=Add predefined workload
Bottom
F12=Cancel
Figure 99. Choosing Option 8 to Select the Required Workloads
g. Save the workload to the base model.
Change the name of the workload to a name with system identification
and save it to the library of the base model.
For example, you can rename the CLIENTAC4 workload as SYSACA4 and
fill in QPFR_TIM53 for the base model’s library (see Figure 100 on
page 193).
192
Slicing the AS/400 with Logical Partitioning: A How to Guide
Save Workload to Workload Member
Change values if desired, press Enter.
Member . . . . . . . . . . . . SYSACA4
Library . . . . . . . . . .
QPFR_TIM53
Name
Name
Text . . . . . . . . . . . . .
Interactive performance data from SG system
Replace . . . . . . . . . . .
N
Y=Yes, N=No
CPU architecture . . . . . . .
*RISC
*CISC, *RISC
F12=Cancel
Figure 100. Save Workload to Workload Member
After you save all of the required workloads, you return to the Work with
Workloads display.
Work with Workloads
Model/Text:
INTER01
Interactive Performance
Type options, press Enter.
1=Create 2=Change 3=Copy 4=Delete
8=Save workload to workload member
Opt
Workload
Text
CLIENTAC4
INTERACTIV
NONINTER
QDEFAULT
Interactive
Interactive
Interactive
Interactive
F3=Exit F6=Add saved workload
F13=Combine workloads
5=Display 6=Print
9=Edit transactions
performance
performance
performance
performance
data
data
data
data
from
from
from
from
SG
SG
SG
SG
7=Rename
system
system
system
system
F9=Add predefined workload
Bottom
F12=Cancel
Figure 101. Work with Workloads
Press F12 to exit.
Interactive and Batch CPW Determination
193
h. Now, you are ready to import the saved workloads to the base model.
Go back to Work with BEST/1 Model and work with the base model,
M53SINT in library QPFR_TIM53.
Choose option 1 (Work with workloads) to work with the base model’s
workloads.
Work with BEST/1 Model
Performance data . . . :
Model/Text . . . . . . :
QPFR_TIM53 (Q983510730)
M53SINT
Select one of the following:
1. Work with workloads
2. Specify objectives and active jobs
5. Analyze current model
6. Analyze current model and give recommendations
7. Specify workload growth and analyze model
10. Configuration menu
11. Work with results
More...
Selection or command
===> 1
F20=More BEST/1 options
F21=Basic user level
F24=More keys
Figure 102. Work with Best/1
i. Press F6 to add the saved workload to the base model (Figure 103 on
page 195).
194
Slicing the AS/400 with Logical Partitioning: A How to Guide
Work with Workloads
Model/Text:
M53SINT
Type options, press Enter.
1=Create 2=Change 3=Copy 4=Delete
8=Save workload to workload member
Opt
Workload
Text
CLIENTAC4
INTERACTIV
NONINTER
QDEFAULT
Interactive
Interactive
Interactive
Interactive
F3=Exit F6=Add saved workload
F13=Combine workloads
5=Display 6=Print
9=Edit transactions
performance
performance
performance
performance
data
data
data
data
from
from
from
from
HK
HK
HK
HK
7=Rename
system
system
system
system
F9=Add predefined workload
Bottom
F12=Cancel
Figure 103. Work with Workloads
j. Select the previous saved workload. Enter a 1 next to the required
workloads for selection and press Enter. You need to select the
workload one by one. See Figure 104.
Add Saved Workload
Library . . . . .
QPFR_TIM53
Name
Type option, press Enter.
1=Select
Opt Member
1 SYSANINTR
SYSAINTR
SYSADFT
SYSACA4
Text
Interactive performance data from SG
Interactive performance data from SG
Interactive performance data from SG
Interactive performance data from SG
F3=Exit F5=Refresh F12=Cancel
F19=Sort by date and time
F15=Sort by member
Date
02/04/99
02/04/99
02/04/99
02/04/99
Time
15:45:28
15:45:19
15:45:13
15:45:05
Bottom
F16=Sort by text
Figure 104. Add Saved Workload
Interactive and Batch CPW Determination
195
The workloads are added successfully.
Work with Workloads
Model/Text:
M53S
Type options, press Enter.
1=Create 2=Change 3=Copy 4=Delete
8=Save workload to workload member
Opt
Workload
Text
CLIENTAC4
INTERACTIV
NONINTER
QDEFAULT
SYSACA4
SYSADFT
SYSAINTR
SYSANINTR
Interactive
Interactive
Interactive
Interactive
Interactive
Interactive
Interactive
Interactive
5=Display 6=Print
9=Edit transactions
performance
performance
performance
performance
performance
performance
performance
performance
data
data
data
data
data
data
data
data
from
from
from
from
from
from
from
from
HK
HK
HK
HK
SG
SG
SG
SG
7=Rename
system
system
system
system
system
system
system
system
Bottom
F3=Exit F6=Add saved workload F9=Add predefined workload F12=Cancel
F13=Combine workloads
Workload SYSACA4 has been read. Check ASP, pool, and comm line assignments
Figure 105. Work with Workload
k. You may need to add additional storage pools, disk units to ASPs, or
communication adapters to the base model to support the combined
workload.
l. You can continue with the normal performance projection by specifying
the growth objective and choosing the user-defined AS/400 partition
model.
Refer to 6.7, “Creating a New Hardware Feature in BEST/1” on page
99, to learn how to create a user-defined AS/400 model for an AS/400
partition.
Once you satisfy the projected performance on the user-defined AS/400
partition model, you can convert the Normal Performance rating of the
user-defined AS/400 partition model back to an interactive CPW number.
196
Slicing the AS/400 with Logical Partitioning: A How to Guide
Work with CPU Models
Type option, press Enter.
1=Create 2=Change 3=Copy
Opt CPU Model
650
730
730
730
730
730
730
730
730
730
730
730
F3=Exit
B10
LP04
2A6A
2A6B
2A6C
2A6D
2A6E
2A6F
2B6A
2B6B
2B6C
2B6D
2B6E
4=Delete
7=Rename
System Archi- --Relative Perf--- Number of Max Stor
Unit tecture
Normal
Server Processors
(MB)
9404
9406
9406
9406
9406
9406
9406
9406
9406
9406
9406
9406
9406
*CISC
*RISC
*RISC
*RISC
*RISC
*RISC
*RISC
*RISC
*RISC
*RISC
*RISC
*RISC
*RISC
1.00
262.35
18.10
29.33
59.02
147.36
17.74
29.33
58.29
145.91
274.80
17.02
54.67
147.36
147.36
147.36
274.80
274.80
274.80
274.80
505.43
505.43
F9=Display equivalent BEST/1 and IBM CPU models
1
4
1
1
1
1
2
2
2
2
2
4
4
16
32768
24576
24576
24576
24576
24576
24576
24576
24576
24576
24576
24576
More...
F12=Cancel
Figure 106. User-Defined AS/400 Partition Model — 650-LP04
For example, the projected performance fits into a user-defined AS/400
partition Model 640-LP04 with interactive CPU utilization around 70% and
interactive response time less than 1.5 seconds. The user-defined AS/400
partition model 640-LP04 is created by using the Model 9406-640 #2240
8-way processor with a relative performance rating of 524.70 as the base.
However, the Model 640-LP04 has only four processors and a relative
performance rating of 262.35. Since the interactive CPW of a 9406-640
#2240 processor is 1,794, the required interactive for the combined workload
is:
1,794 x (262.35 / 524.70) = 897
E.3 Batch Performance Determination
You should have performed step 1 through step 5 in E.1, “Overview of the
Methodology” on page 189, before you continue to determine the batch
performance requirement. Now you have a clear idea of your current
performance level and whether the current performance level matches the
expectation.
For example, a customer has two AS/400 servers running the same ERP
application. One is located in Hong Kong and the other is located in
Interactive and Batch CPW Determination
197
Singapore. Both AS/400 servers have the same capacity, 9406-53S #2157
processor with 2 GB of main storage. The system loading for both places is
roughly the same for the batch processing at night. The batch processing is
normally started at 7:30pm and ended at 11:30pm. Figure 107 shows the
CPU utilization versus time chart of a typical batch processing of the
customer.
Figure 107. CPU Utilization versus Time Chart
The customer wants to consolidate the two 9406-53S servers into one of the
two partitions of a new Model 9406-S40 #2207 8-way processor. The other
partition is used for the new workflow application running on Domino/400.
What is the required batch CPW for the combined workload?
Let us look at one system first. Figure 108 on page 199 shows that the
9406-53S has been driven up to 98.4% and the disk arm utilization is low,
only 2%. Therefore, disk arm contention is not the bottleneck. To complete
the batch processing within four hours, you need an AS/400 server with
comparable batch performance. The client/server CPW (or batch CPW) of the
9406-53S #2157 processor is 650. Therefore, you need a partition with at
least 1,300 (650 x 2) batch CPW.
198
Slicing the AS/400 with Logical Partitioning: A How to Guide
Display Analysis Summary
Period
Base
+10%
+10%
+10%
+10%
CPU
53S
53S
53S
53S
53S
Model
2157
2157
2157
2157
2157
Period
Base
+10%
+10%
+10%
+10%
----Inter
Local
.5
.5
.5
.5
.5
Stor
(MB)
2048
2048
2048
2048
2048
CPU -Disk IOPs-- -Disk Ctls-- -Disk Arms-Util Nbr
Util Nbr
Util Nbr
Util
98.4
5
5.6
11
.8
70
2.2
99.0
5
5.8
11
.9
70
2.3
99.0
5
6.2
11
.9
70
2.5
99.0
5
6.7
11
1.0
70
2.7
99.0
5
7.3
11
1.1
70
3.0
Bottom
Rsp Time---- -------Inter-------- -----Non-Inter-----LAN
WAN CPU Util
Trans/Hr CPU Util
Trans/Hr
4.1
.0
.0
47
98.4
14032018
4.1
.0
.0
52
99.0
13898859
4.1
.0
.0
57
99.0
13629150
4.1
.0
.1
63
98.9
13332477
4.1
.0
.1
69
98.9
13006129
Bottom
F3=Exit F10=Re-analyze
F15=Configuration menu
F11=Alternative view
F17=Analyze multiple points
F12=Cancel
F24=More keys
Figure 108. BEST/1 Projection for the Batch Processing
To project the performance requirement, you need to save the workloads from
the BEST/1 model for one of the Model 9406-53S servers. Please refer to
E.2, “Interactive Performance Determination” on page 190, for details.
We assume that you saved workloads to the base model, M53S and added to
the workload profile as shown in Figure 109 on page 200.
Interactive and Batch CPW Determination
199
Work with Workloads
Model/Text:
M53S
Type options, press Enter.
1=Create 2=Change 3=Copy 4=Delete
8=Save workload to workload member
Opt
Workload
Text
CLIENTAC4
INTERACTIV
NONINTER
QDEFAULT
SYSBCA4
SYSBDFT
SYSBINTR
SYSBNINTR
Batch
Batch
Batch
Batch
Batch
Batch
Batch
Batch
performance
performance
performance
performance
performance
performance
performance
performance
F3=Exit F6=Add saved workload
F13=Combine workloads
5=Display 6=Print
9=Edit transactions
data
data
data
data
data
data
data
data
for
for
for
for
for
for
for
for
HK
HK
HK
HK
SG
SG
SG
SG
7=Rename
system
system
system
system
system
system
system
system
F9=Add predefined workload
Bottom
F12=Cancel
Figure 109. Workload Profile of M53S
Now, go to Work with BEST/1 Model display and select option 10
(Configuration menu).
As shown in Figure 110 on page 201, select option 1 (Change CPU and other
resource values) and press Enter.
200
Slicing the AS/400 with Logical Partitioning: A How to Guide
Configuration
CPU Model . . .
Main stor (MB) .
Main stor pools
Disk IOPs . . .
Disk ctls . . .
Disk arms . . .
ASPs . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
:
:
:
:
:
:
:
2157
2048
3
5
11
70
1
Comm IOPs . . . . . . . . :
LAN lines . . . . . . . . :
WAN lines . . . . . . . . :
1
1
0
Multifunction IOPs
Disk IOAs . . . .
Comm IOAs . . . .
IPCS IOAs . . . .
1
0
0
0
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
:
:
:
:
Select one of the following:
1.
2.
3.
4.
5.
Change CPU and other resource values
Work with disk resources
Edit ASPs
Edit main storage pools
Work with communications resources
Selection or command
===> 1
F3=Exit F4=Prompt F9=Retrieve
F17=Correct configuration
F12=Cancel F13=Check configuration
F24=More keys
Figure 110. Configuration
Select the user defined AS/400 partition model, S40-LP01 (4 processors),
based on 9406-S40 #2207 8-way server using option 1.
Select CPU
Type option, press Enter.
1=Select
Opt CPU Model
B10
1 S40 LP01
20S 2010
200 2030
200 2031
200 2032
300 2040
300 2041
300 2042
310 2043
310 2044
320 2050
320 2051
System Archi- --Relative Perf--- Number of Max Stor
Unit tecture
Normal
Server Processors
(MB)
9404
*CISC
1.00
1
16
9406
*RISC
15.00
500.00
4
40960
9402
*CISC
1.90
5.90
1
128
9402
*CISC
2.50
1
24
9402
*CISC
4.00
1
56
9402
*CISC
5.80
1
128
9406
*CISC
4.00
1
72
9406
*CISC
5.80
1
80
9406
*CISC
7.20
1
160
9406
*CISC
10.90
1
832
9406
*CISC
19.00
2
832
9406
*CISC
23.20
1
1536
9406
*CISC
41.30
2
1536
More...
F12=Cancel
Figure 111. Select CPU
Interactive and Batch CPW Determination
201
The S40-LP01 is created from the AS/400 server, Model 9406-S40 #2207
8-way processor. The S40-LP01 has only four processors with a relative
normal performance (interactive) rating of 15 and relative server performance
rating of 500.
CPU
Model
System
Unit
S40
2207
9406
Archi- --Relative Perf--Number of Max Stor
tecture
Normal
Server Processors
(MB)
*RISC
30.41
998.55
8
40960
Note
BEST/1 uses its own performance rating figure. It is not the CPW number.
You can project the combined workloads on the user defined model,
S40-LP01. The result is shown in Figure 112.
Display Analysis Summary
Period
Base
+10%
+10%
+10%
+5%
CPU
S40
S40
S40
S40
S40
Model
LP01
LP01
LP01
LP01
LP01
Period
Base
+10%
+10%
+10%
+5%
----Inter
Local
.7
.7
.7
.7
.7
Stor
(MB)
2048
2048
2048
2048
2048
CPU -Disk IOPs-- -Disk Ctls-- -Disk Arms-Util Nbr
Util Nbr
Util Nbr
Util
73.5
5
13.5
9
2.4
70
5.2
80.8
5
14.9
9
2.7
70
5.8
88.9
5
16.4
9
2.9
70
6.4
97.8
5
18.0
9
3.2
70
7.2
99.0
5
18.1
9
3.2
70
7.2
Bottom
Rsp Time---- -------Inter-------- -----Non-Inter-----LAN
WAN CPU Util
Trans/Hr CPU Util
Trans/Hr
4.3
.0
.0
95
73.4
28064038
4.3
.0
.0
104
80.8
30870438
4.3
.0
.0
114
88.9
33957484
4.4
.0
.0
126
97.8
37353236
4.3
.0
.0
132
99.0
37619880
Bottom
F3=Exit F10=Re-analyze
F15=Configuration menu
F11=Alternative view
F17=Analyze multiple points
F12=Cancel
F24=More keys
Figure 112. Projected Batch Performance for the Combined Workload
To simplify the situation, we made the assumption that the workloads on the
two machines are exactly similar and each one of them drives the 9406-53S
#2157 4-way processor to 98.4% with 14,032,018 transactions per hour. This
allows us to directly combine the batch runtimes of the two workloads on the
partition with four processors of a 9406-S40 #2207 8-way server.
202
Slicing the AS/400 with Logical Partitioning: A How to Guide
The projected runtime of the combined workload is:
4 hours x ((14,032,018 + 14,032,018)/ 37,619,880) = 3 hours.
The batch CPW of the S40-LP01 is:
3660 x (500 / 998.55) = 1833.
Where 3660 is the 9406-S40 #2207 processor’s C/S CPW.
If the customer accepts the current batch window, four hours, then the
required batch performance for the combined workload is 1,300 (650 x2)
batch CPW. However, if the customer wants to shorten the batch window from
four hours to three hours, then the required batch performance is 1,833 batch
CPW.
Interactive and Batch CPW Determination
203
204
Slicing the AS/400 with Logical Partitioning: A How to Guide
Appendix F. Special Notices
This publication is intended to help IBM Business Partners and IBM Global
Service Professionals to prepare and plan for implementing Logical
Partitioning on AS/400. The information in this publication is not intended as
the specification of any programming interfaces that are provided by the
AS/400 system. See the PUBLICATIONS section of the IBM Programming
Announcement for the AS/400 for more information about what publications
are considered to be product documentation.
References in this publication to IBM products, programs or services do not
imply that IBM intends to make these available in all countries in which IBM
operates. Any reference to an IBM product, program, or service is not
intended to state or imply that only IBM's product, program, or service may be
used. Any functionally equivalent program that does not infringe any of IBM's
intellectual property rights may be used instead of the IBM product, program
or service.
Information in this book was developed in conjunction with use of the
equipment specified, and is limited in application to those specific hardware
and software products and levels.
IBM may have patents or pending patent applications covering subject matter
in this document. The furnishing of this document does not give you any
license to these patents. You can send license inquiries, in writing, to the IBM
Director of Licensing, IBM Corporation, 500 Columbus Avenue, Thornwood,
NY 10594 USA.
Licensees of this program who wish to have information about it for the
purpose of enabling: (i) the exchange of information between independently
created programs and other programs (including this one) and (ii) the mutual
use of the information which has been exchanged, should contact IBM
Corporation, Dept. 600A, Mail Drop 1329, Somers, NY 10589 USA.
Such information may be available, subject to appropriate terms and
conditions, including in some cases, payment of a fee.
The information contained in this document has not been submitted to any
formal IBM test and is distributed AS IS. The information about non-IBM
("vendor") products in this manual has been supplied by the vendor and IBM
assumes no responsibility for its accuracy or completeness. The use of this
information or the implementation of any of these techniques is a customer
responsibility and depends on the customer's ability to evaluate and integrate
them into the customer's operational environment. While each item may have
© Copyright IBM Corp. 1999
205
been reviewed by IBM for accuracy in a specific situation, there is no
guarantee that the same or similar results will be obtained elsewhere.
Customers attempting to adapt these techniques to their own environments
do so at their own risk.
Any pointers in this publication to external Web sites are provided for
convenience only and do not in any manner serve as an endorsement of
these Web sites.
Any performance data contained in this document was determined in a
controlled environment, and therefore, the results that may be obtained in
other operating environments may vary significantly. Users of this document
should verify the applicable data for their specific environment.
The following document contains examples of data and reports used in daily
business operations. To illustrate them as completely as possible, the
examples contain the names of individuals, companies, brands, and products.
All of these names are fictitious and any similarity to the names and
addresses used by an actual business enterprise is entirely coincidental.
Reference to PTF numbers that have not been released through the normal
distribution process does not imply general availability. The purpose of
including these reference numbers is to alert IBM customers to specific
information relative to the implementation of the PTF when it becomes
available to each customer according to the normal IBM PTF distribution
process.
The following terms are trademarks of the International Business Machines
Corporation in the United States and/or other countries:
CT
DRDA
IBM
Netfinity
OS/2
RS/6000
Service Director
System/38
VisualAge
XT
â
206
DB2
GDDM
ImagePlus
OfficeVision
OS/400
S/370
SP
System/390
VisualInfo
400
Slicing the AS/400 with Logical Partitioning: A How to Guide
The following terms are trademarks of other companies:
C-bus is a trademark of Corollary, Inc. in the United States and/or other
countries.
Java and all Java-based trademarks and logos are trademarks or registered
trademarks of Sun Microsystems, Inc. in the United States and/or other
countries.
Microsoft, Windows, Windows NT, and the Windows logo are trademarks of
Microsoft Corporation in the United States and/or other countries.
PC Direct is a trademark of Ziff Communications Company in the United
States and/or other countries and is used by IBM Corporation under license.
ActionMedia, LANDesk, MMX, Pentium and ProShare are trademarks of Intel
Corporation in the United States and/or other countries.
UNIX is a registered trademark in the United States and/or other countries
licensed exclusively through X/Open Company Limited.
SET and the SET logo are trademarks owned by SET Secure Electronic
Transaction LLC.
Other company, product, and service names may be trademarks or service
marks of others.
Special Notices
207
208
Slicing the AS/400 with Logical Partitioning: A How to Guide
Appendix G. Related Publications
The publications listed in this section are considered particularly suitable for a
more detailed discussion of the topics covered in this redbook.
G.1 International Technical Support Organization Publications
For information on ordering these ITSO publications see “How to Get ITSO
Redbooks” on page 211.
• AS/400e System Handbook V4R4 , GA19-5486
• AS/400 System Builder V4R4 , SG24-2155
• AS/400 Server Capacity Planning , SG24-2159
• AS/400 Consolidation Strategies and Implementation, SG24-5186
• AS/400 Remote Journal Function for High Availability and Data
Replication, SG24-5189
G.2 Redbooks on CD-ROMs
Redbooks are also available on the following CD-ROMs. Click the CD-ROMs
button at http://www.redbooks.ibm.com/ for information about all the CD-ROMs
offered, updates and formats.
CD-ROM Title
System/390 Redbooks Collection
Networking and Systems Management Redbooks Collection
Transaction Processing and Data Management Redbooks Collection
Lotus Redbooks Collection
Tivoli Redbooks Collection
AS/400 Redbooks Collection
Netfinity Hardware and Software Redbooks Collection
RS/6000 Redbooks Collection (BkMgr)
RS/6000 Redbooks Collection (PDF Format)
Application Development Redbooks Collection
© Copyright IBM Corp. 1999
Collection Kit
Number
SK2T-2177
SK2T-6022
SK2T-8038
SK2T-8039
SK2T-8044
SK2T-2849
SK2T-8046
SK2T-8040
SK2T-8043
SK2T-8037
209
G.3 Other Publications
These publications are also relevant as further information sources:
• AS/400 Road Map for Changing to PowerPC Technology V4R3 ,
SA41-5150
• System Upgrade Roadmap (RISC to RISC) V4R3, SA41-5155
• OS/400 Distributed Data Management , SC41-3307
• Client Access for Windows 95/NT Setup V3R2, SC41-3512
• Basic System Operation, Administration, and Problem Handling,
SC41-5206
• Backup and Recovery, SC41-5304
• Performance Tools for AS/400, SC41-5340
• BEST/1 Capacity Planning Tool V4R2, SC41-5341
• OptiConnect for OS/400 , SC41-5414
• Operations Console Setup, SC41-5508
• AS/400 System API Reference V4R3 , SC41-5801
210
Slicing the AS/400 with Logical Partitioning: A How to Guide
How to Get ITSO Redbooks
This section explains how both customers and IBM employees can find out about ITSO redbooks,
redpieces, and CD-ROMs. A form for ordering books and CD-ROMs by fax or e-mail is also provided.
• Redbooks Web Site http://www.redbooks.ibm.com/
Search for, view, download, or order hardcopy/CD-ROM redbooks from the redbooks Web site. Also
read redpieces and download additional materials (code samples or diskette/CD-ROM images) from
this redbooks site.
Redpieces are redbooks in progress; not all redbooks become redpieces and sometimes just a few
chapters will be published this way. The intent is to get the information out much quicker than the
formal publishing process allows.
• E-mail Orders
Send orders by e-mail including information from the redbooks fax order form to:
In United States
Outside North America
e-mail address
[email protected]
Contact information is in the “How to Order” section at this site:
http://www.elink.ibmlink.ibm.com/pbl/pbl/
• Telephone Orders
United States (toll free)
Canada (toll free)
Outside North America
1-800-879-2755
1-800-IBM-4YOU
Country coordinator phone number is in the “How to Order”
section at this site:
http://www.elink.ibmlink.ibm.com/pbl/pbl/
• Fax Orders
United States (toll free)
Canada
Outside North America
1-800-445-9269
1-403-267-4455
Fax phone number is in the “How to Order” section at this site:
http://www.elink.ibmlink.ibm.com/pbl/pbl/
This information was current at the time of publication, but is continually subject to change. The latest
information may be found at the redbooks Web site.
IBM Intranet for Employees
IBM employees may register for information on workshops, residencies, and redbooks by accessing
the IBM Intranet Web site at http://w3.itso.ibm.com/ and clicking the ITSO Mailing List button.
Look in the Materials repository for workshops, presentations, papers, and Web pages developed
and written by the ITSO technical professionals; click the Additional Materials button. Employees
may access MyNews at http://w3.ibm.com/ for redbook, residency, and workshop announcements.
© Copyright IBM Corp. 1999
211
IBM Redbook Fax Order Form
Please send me the following:
Title
Order Number
First name
Quantity
Last name
Company
Address
City
Postal code
Country
Telephone number
Telefax number
VAT number
Card issued to
Signature
Invoice to customer number
Credit card number
Credit card expiration date
We accept American Express, Diners, Eurocard, Master Card, and Visa. Payment by credit card not
available in all countries. Signature mandatory for credit card payment.
212
Slicing the AS/400 with Logical Partitioning: A How to Guide
Index
Symbols
$DATE 126
$INTVL 126
$MON 126
$RDUR 126
$RVAL 126
$SEQ 126
$TDUR 126
$TIME 126
$TVAL 126
$VAL 126
Numerics
24 x 7 database availability 164
A
accepting load source disk unit 75
accessing LPAR functions 175
adding logical partitions 111
addition 170
ALCOBJ (Allocate Object) command 136
Allocate Object (ALCOBJ) command 136
application stability 19, 28
application testing 29
archiving solution 4
AS/400
custom server models 87
server models 86
system models 83
AS/400 Logical Partitioning 1
AS/400 workload 82
AS/400e server 7XX 84
assigning CPW for batch 91
auditing end action (QAUDENDACN) 49
automatic IPL after power restored (QPWRRSTIPL)
47
availability requirements 18, 28
B
backup and recovery, minimizing windows 6
batch capacity planning for LPAR 89
batch CPW 88
batch CPW determination 189
batch performance determination 197
BATCH400 89
BEST/1 99
creating a new hardware feature 99
creating a new hardware features 103
bus 39, 167
bus-level partitions 29
C
C program sample 173
capacity planning 81
capacity planning tools 98
Change IPL Attributes (CHGIPLA) command 50
change management 20, 63
changing partition processing resources 44
changing the resources 44
Check Object (CHKOBJ) command 136
CHGIPLA (Change IPL Attributes) command 50
CHKOBJ (Check Object) command 136
clearing configuration data 76
client command 126
client command parameters 126
clustering 13
collecting performance data 113
collection interval 121
collection services
on a single partition 113
on a system group 114
command definition 127
Commercial Processing Workload (CPW) 82
communication arbiters (QCMNARB) 49
configuration 21
configuration data
clearing 76
copying 76
updating 75
configuration data for logical partitions 73
configuration management 21
configurations 73
console 169
consolidated performance tools 98
consolidating worldwide operations 7
copying configuration data 76
CPW (Commercial Processing Workload) 82
CPW for batch 91
CPW values 83
creating a new hardware feature in BEST/1 99
creating a new monitor 119
creating system groups 111
213
custom server models 87
customer requirements 17
capacity planning and performance 81
D
DASD compression support 37
data replication
using journals 160
using the remote journal function 161
database files 116
Date and time to automatically IPL (QIPLDATTIM)
50
date processing 51
DB2 Multisystem 165
DDM (Distributed Data Management) 135
dedicated mode 169
dedicated service tools (DST) 27
determining batch CPW 88
directory entry 158
Display OptiConnect Link Status (DSPOPCLNK)
command 152
display time 122
Distributed Data Management (DDM) 135
distributing changes through Management Central
64
distribution queue 159
Domino clustering 9
DSPOPCLNK (Display OptiConnect Link Status)
command 152
DST (dedicated service tools) 27
DST hardware resource information 53
dual tape attach support 37
duration 125
I
I/O resources temporarily unavailable 43
IBM code testing 29
ILE RPG program sample 171
implementing an archiving solution 4
input/output adapter (IOA) 167
input/output processor (IOP) 167
installing V4R4M0 25
interactive CPW determination 189
interactive performance 39, 96
temporarily unavailable 42
interactive performance determination 190
interactive workload 82
inter-partition communication 133
INTNETADR (Internet address) 154
IOA (input/output adapter) 167
IOP 40
moving existing 43
IOP (input/output processor) 167
IOP-level partitions 29
IPL process 59
J
E
journal, data replication 160
Electronic Customer Support line 72
expansion unit 169
Extended Function Path Routing 144
F
Fastpath OptiConnect 142
H
HABP (High Availability Business Partner) applications 34
hardware
considerations 23
214
limits 37
making it temporarily unavailable 40
hardware feature, creating in BEST/1 99
hardware features in BEST/1 103
hardware resource configuration 53
High Availability Business Partner (HABP) applications 34
host command 125
host parameters 126
Hypervisor 168
L
LCLIFC (associated local interface) 154
LCS (Local Controlling System) 178
license key management 59
Licensed Internal Code Log 72
licensed program products (LPP) 24
licensing LPP 24
LIND (line description) 154
load source 168
Local Controlling System (LCS) 178
log, Licensed Internal Code 72
Slicing the AS/400 with Logical Partitioning: A How to Guide
logical partition
configurations 73
managing security 68
managing user applications 63
power interruptions 72
problem determination 69
problem management 69
recovering user data 78
restoring user data 78
logical partitioning 1
application dependencies 26
batch capacity planning 89
change management 63
concept view 12
configuration 21
date and time processing 51
dedicated mode 169
environment 31
history 1
how it works 11
installing new hardware 39
planning 17
saving and restoring 45
saving the configuration 46
scenario 2
shared mode 169
software problems 25
supported models 23
system values affected 47
terminology 11, 167
logical partitions
adding 111
calculating new performance ratings 102
configuration data 73
hardware resource information 55
recovery 77
removing hardware resources 57
transferring data 140
LPAR functions 175
LPP (licensed program products) 24
M
Machine Interface (MI) 161
machine pool 60
main storage 39, 95, 167
temporarily unavailable 41
main storage dump processing 52
Management Central 118
distributing changes 64
for performance management 110
in PTF distribution 66
managing fixes 67
managing fixes with Management Central 67
managing partitions 63
managing security in logical partition 68
managing user applications in a logical partition 63
marketplace positioning 15
maximum graphing value 122
MFIOP (multifunction IOP) 168
MI (Machine Interface) 161
minimizing backup and recovery windows 6
mixed production environment 3
monitor graphs 129
monitor, running 128
monitoring realtime system performance 118
moving existing IOPs 43
moving to a partitioned environment 31
multidatabase applications 6
multifunction IOP (MFIOP) 168
multiple production environment 2
N
naming convention, system 31
national language considerations 26
new monitor 119
nonconfigured disk unit configuration data 75
non-interactive workload 82
nonreporting logical partitioning resources 75
number of processors 93
O
ObjectConnect 35
ObjectConnect/400 137
operational considerations 39
Operations Console 177
connection for a partitioned AS/400 177
in secondary partition 178
operations management 20
OptiConnect 170
interface 153
OptiConnect/400 133
bandwidth 134
capabilities 133
setting up over TCP/IP 152
OptiConnect/400 performance 97
OptiMover for OS/400 135
215
order processing 21
OS/400 functions using OptiConnect/400 135
Override with Database File (OVRDBF) command
136
OVRDBF (Override with Database File) command
136
ownership 169
P
PAL (Product Activity Log) 70
parameter 126
Partition Licensed Internal Code (PLIC) 168
partition processing 44
partition processing configuration, getting there
100
partition processing resources 92
partition security 29
partitions
bus-level 29
IOP-level 29
performance 81
performance data, collecting 113
performance management 20
using Management Central 110
Performance Management/400 109
performance monitoring 130
performance ratings for logical partitions 102
Performance Tools 99
planning 17
PLIC (Partition Licensed Internal Code) 168
PM/400 109
points of failure 17
power interruptions 72
Preventive Service Planning (PSP) 72
primary partition 12
available information 55
recovering configuration data 74
printing the system configuration 58
problem determination for logical partition 69
problem management 20
problem management for logical partition 69
processor 167
processor feature (QPRCFEAT) 50
processor multi-tasking (QPRCMLTTSK) 47
processors 39
temporarily unavailable 40
Product Activity Log (PAL) 70
PSP (Preventive Service Planning) 72
216
PTF distribution, using Management Central 66
PTF management 65
Q
QAUDENDACN value 49
QCMNARB value 49
QIPLDATTIM value 50
QMODEL value 50
QPFRADJ value 61
QPRCFEAT value 50
QPRCMLTTSK value 47
QPWRRSTIPL value 47
QRMTIPL value 48
QSRLNBR value 50
Query/400 137
QUPSDLYTIM value 48
R
recovering user data from a deleted logical partition
78
recovery
considerations with logical partitions 77
for logical partitions 77
primary partition configuration data 74
system initialization 78
related system commands 49
related system values 49
Remote Controlling System (RSC) 178
remote data areas 137
remote data queues 137
remote journal function 138, 159
benefits 164
data replication 161
setup 162
remote power on and IPL (QRMTIPL), QRMTIPL
value 48
removable media device 169
removal 170
reporting to IBM 70
requirements
application stability 19
availability 18, 28
customer 17
resource 19
security 20
reset value 122
resource allocation 29
resource requirements 19
Slicing the AS/400 with Logical Partitioning: A How to Guide
restoring data to logical partitions 46
restoring partitions 45
restoring user data to a logical partition 78
retention period 121
retrieving resources allocated to a partition 44
routing table 159
RSC (Remote Controlling System) 178
S
SAL (Service Action Log) 70
sample C program 173
sample ILE RPG program 171
save strategy 33
Save/Restore (SAVRST) command 137
Save/Restore Changed Objects (SAVRSTCHG)
command 137
Save/Restore Configuration (SAVRSTCFG) command 138
Save/Restore Document Library Object (SAVRSTDLO) command 138
Save/Restore Library (SAVRSTLIB) command 138
Save/Restore Object (SAVRSTOBJ) command
137
saving data from logical partitions 46
saving partitions 45
saving the configuration 46
SAVRST (Save/Restore) command 137
SAVRSTCFG (Save/Restore Configuration) command 138
SAVRSTCHG (Save/Restore Changed Objects)
command 137
SAVRSTDLO (Save/Restore Document Library Object) command 138
SAVRSTLIB (Save/Restore Library) command 138
SAVRSTOBJ (Save/Restore Object) command
137
SBMRMTCMD (Submit Remote) command 137
secondary partition 12
available information 57
security 32
security considerations 51
security requirements 20
serial number, system 32
server models 86
Service Action Log (SAL) 70
service provider 79
setting up remote journal function 162
setting up the central system 110
shared mode 169
shared nothing 165
sharing 169
single partition, starting collection services 113
SLIC (System Licensed Internal Code) 168
SNA Distribution Services (SNADS) 158
software considerations 24
SQL (Standard Query Language) 137
SRC (System Reference Codes) 70
SST (system service tools) 27
hardware resource information 53
Standard Query Language (SQL) 137
Submit Remote (SBMRMTCMD) command 137
SUBNETMASK (subnet mask) 154
system configuration, printing 58
system consolidation 5
system group, starting collection services 114
system groups 111
system initialization recovery 78
System Licensed Internal Code (SLIC) 168
system model number (QMODEL) 50
system models 83
system naming convention 31
System Reference Codes (SRC) 70
system serial number 32
system serial number (QSRLNBR) 50
system service tools (SST) 27
system values affected by logical partitioning 47
T
tape drives 33
tape library 33
TCP/IP, setting up OptiConnect/400 152
test environment 3
three-tiered application architecture 10
threshold commands 126
threshold reset 124
threshold trigger 124
options 122
time processing 51
trigger value 122
U
uninterruptible power supply delay time (QUPSDLYTIM) 48
updating configuration data 75
217
V
value 125
Virtual OptiConnect
activity 149
ending 148
starting 148
Virtual OptiConnect/400 139
application examples 158
W
work management 60
Work with Active Jobs (WRKACTJOB) command
149
Work with OptiConnect Activity (WRKOPCACT)
command 150
workload 82
interactive 82
non-interactive 82
worldwide operations, consolidating 7
WRKACTJOB (Work with Active Jobs) command
149
WRKOPCACT (Work with OptiConnect Activity)
command 150
218
Slicing the AS/400 with Logical Partitioning: A How to Guide
ITSO Redbook Evaluation
Slicing the AS/400 with Logical Partitioning: A How to Guide
SG24-5439-00
Your feedback is very important to help us maintain the quality of ITSO redbooks. Please complete this
questionnaire and return it using one of the following methods:
• Use the online evaluation form found at http://www.redbooks.ibm.com/
• Fax this form to: USA International Access Code + 1 914 432 8264
• Send your comments in an Internet note to [email protected]
Which of the following best describes you?
_ Customer _ Business Partner
_ Solution Developer
_ None of the above
_ IBM employee
Please rate your overall satisfaction with this book using the scale:
(1 = very good, 2 = good, 3 = average, 4 = poor, 5 = very poor)
Overall Satisfaction
__________
Please answer the following questions:
Was this redbook published in time for your needs?
Yes___ No___
If no, please explain:
What other redbooks would you like to see published?
Comments/Suggestions:
© Copyright IBM Corp. 1999
(THANK YOU FOR YOUR FEEDBACK!)
219
Slicing the AS/400 with Logical Partitioning: A How to Guide
SG24-5439-00
Printed in the U.S.A.
SG24-5439-00